HK40023655B - Method for processing medical images, method and device for processing images - Google Patents

Description

A method for medical image processing, an image processing method and apparatus

Technical Field

This application relates to the field of artificial intelligence, and in particular to a method for medical image processing, an image processing method and apparatus.

Background Technology

With the development of medical technology, the recognition and analysis of whole-slide images (WSI) have played an important role in the medical field. Since WSI images typically have sides of tens of thousands of pixels, they need to be scaled or segmented into smaller images for analysis. In this process, most of the background area of the WSI image needs to be removed, while the area containing pathological tissue sections is segmented for subsequent image analysis.

Currently, the main method for extracting pathological tissue regions from WSI images is to first reduce the WSI image to a certain scale and then convert it into a grayscale image. Then, further image processing is performed on the grayscale image, such as image binarization and hole removal. Finally, the pathological tissue regions are extracted from the processed image.

However, after scaling, directly converting a color image to a grayscale image will result in the loss of color information, which is an important image feature. Therefore, the extracted pathological tissue areas will not be accurate enough, which can easily lead to deviations in subsequent image analysis.

Summary of the Invention

This application provides a medical image processing method, image processing method and apparatus, which generates a difference image using color information from different channels before binarizing the image, thereby effectively utilizing the color information in the image. The pathological tissue region extracted based on the difference image is more accurate and has a positive impact on subsequent image analysis.

In view of the above, the first aspect of this application provides a method for medical image processing, comprising:

Acquire a medical image to be processed, wherein the medical image to be processed is a color image, and the medical image to be processed includes first image data, second image data and third image data, and the first image data, second image data and third image data respectively correspond to color information under different attributes;

A difference image is generated based on the first image data, the second image data, and the third image data included in the medical image to be processed.

The difference image is binarized to obtain a binary image;

Generate the foreground segmentation result corresponding to the medical image to be processed based on the binarized image.

A second aspect of this application provides an image processing method, comprising:

Acquire a first image to be processed and a second image to be processed, wherein the first image to be processed is a color image and includes first image data, second image data and third image data, and the first image data, second image data and third image data respectively correspond to color information under different channels;

A difference image is generated based on the first image data, the second image data, and the third image data included in the first image to be processed.

The difference image is binarized to obtain a binary image;

Generate the foreground segmentation result corresponding to the first image to be processed based on the binarized image;

Based on the foreground segmentation results, the target object is extracted from the first image to be processed;

A composite image is generated based on the target object and the second image to be processed, wherein the target object is located in the first layer, the second image to be processed is located in the second layer, and the first layer overlaps the second layer.

A third aspect of this application provides a medical image processing apparatus, comprising:

The acquisition module is used to acquire the medical image to be processed, wherein the medical image to be processed is a color image, and the medical image to be processed includes first image data, second image data and third image data, and the first image data, second image data and third image data respectively correspond to color information under different channels;

The generation module is used to generate a difference image based on the first image data, the second image data, and the third image data included in the medical image to be processed;

The processing module is used to binarize the difference image to obtain a binary image;

The generation module is also used to generate the foreground segmentation result corresponding to the medical image to be processed based on the binarized image.

In one possible design, in one implementation of the third aspect of the embodiments of this application,

The generation module is specifically used to generate a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed.

A minimum value image is generated based on the first image data, the second image data, and the third image data included in the medical image to be processed.

Generate a difference image based on the maximum and minimum value images.

In one possible design, in another implementation of the third aspect of the embodiments of this application,

The generation module is specifically used to determine the maximum pixel value corresponding to the target pixel from the first pixel value, the second pixel value, and the third pixel value included in the first image data, the second image data, and the third image data in the medical image to be processed. The first pixel value is the pixel value corresponding to the first pixel position in the first image data, the second pixel value is the pixel value corresponding to the second pixel position in the second image data, the third pixel value is the pixel value corresponding to the third pixel position in the third image data, and the target pixel is the pixel value corresponding to the fourth pixel position in the maximum value image. The first pixel position, the second pixel position, the third pixel position, and the fourth pixel position all correspond to the position of the same pixel in the medical image to be processed.

The generation module is specifically used to determine the minimum pixel value corresponding to the target pixel from the first pixel value, the second pixel value, and the third pixel value based on the first image data, the second image data, and the third image data included in the medical image to be processed.

The generation module is specifically used to subtract the maximum pixel value corresponding to the target pixel in the maximum value image from the minimum pixel value corresponding to the target pixel in the maximum value image, so as to obtain the difference pixel value corresponding to the target pixel in the difference image.

In one possible design, in another implementation of the third aspect of the embodiments of this application,

The generation module is specifically used to generate a difference image to be processed based on the first image data, the second image data, and the third image data included in the medical image to be processed.

Gaussian blur is applied to the difference image to obtain the difference image.

In one possible design, in another implementation of the third aspect of the embodiments of this application, the medical image processing apparatus further includes a determining module;

The determination module is used to determine the segmentation threshold based on the difference image;

The determination module is also used to determine the pixel as the foreground pixel of the binarized image if the pixel value corresponding to the pixel in the difference image is greater than or equal to the segmentation threshold.

The determination module is also used to determine the pixel as the background pixel of the binarized image if the pixel value corresponding to the pixel in the difference image is less than the segmentation threshold.

In one possible design, in another implementation of the third aspect of the embodiments of this application,

The determination module is specifically used to obtain N pixel values corresponding to N pixels based on the difference image, where there is a one-to-one correspondence between pixel values and pixel points, and N is an integer greater than 1.

Determine the pixel value to be processed from N pixel values, where the pixel value to be processed is the maximum value among the N pixel values;

The segmentation threshold is calculated based on the pixel value to be processed and the ratio threshold.

In one possible design, in another implementation of the third aspect of the embodiments of this application,

The generation module is specifically used to detect the background region in the binarized image using a flooding algorithm, wherein the background region includes multiple background pixels;

Based on the binarized image and the background region in the binarized image, obtain the background pixels within the foreground region of the binarized image, wherein the foreground region includes multiple foreground pixels;

By replacing background pixels in the foreground region of a binarized image with foreground pixels, a hole-filling image is obtained.

Median filtering is applied to the hole-filling image to obtain the foreground segmentation result corresponding to the medical image to be processed.

In one possible design, in another implementation of the third aspect of the embodiments of this application,

The processing module is specifically used to perform median filtering on the hole-filling image to obtain a filtered image, wherein the filtered image includes the foreground region to be processed;

Obtain the boundary line of the foreground region to be processed, where the boundary line consists of M pixels, M being an integer greater than 1;

For each of the M pixels on the boundary line, extend outward by K pixels to obtain the foreground segmentation result, where K is an integer greater than or equal to 1.

In one possible design, in another implementation of the third aspect of the embodiments of this application,

The acquisition module is specifically used to acquire raw medical images;

A sliding window method is used to extract medical sub-images from the original medical image;

If a medical sub-image is detected to contain a pathological tissue region, it is identified as a medical image to be processed.

If a medical sub-image is found to not include a pathological tissue region, the medical sub-image is identified as a background image and removed.

In one possible design, in another implementation of the third aspect of the embodiments of this application, the image processing apparatus further includes a training module;

The generation module is also used to generate target positive sample images based on the foreground segmentation results, wherein the target positive sample image belongs to a positive sample image in the positive sample set, and each positive sample image contains a pathological tissue region;

The acquisition module is also used to acquire a negative sample set, wherein the negative sample set includes at least one negative sample image, and each negative sample image does not contain a pathological tissue region;

The training module is used to train the image segmentation model based on the positive and negative sample sets.

A fourth aspect of this application provides an image processing apparatus, comprising:

The acquisition module is used to acquire a first image to be processed and a second image to be processed. The first image to be processed is a color image and includes first image data, second image data and third image data. The first image data, second image data and third image data correspond to color information under different channels.

The generation module is used to generate a difference image based on the first image data, the second image data, and the third image data included in the first image to be processed;

The processing module is used to binarize the difference image to obtain a binary image;

The generation module is also used to generate the foreground segmentation result corresponding to the first image to be processed based on the binarized image;

The extraction module is used to extract the target object from the first image to be processed based on the foreground segmentation results;

The generation module is also used to generate a composite image based on the target object and the second image to be processed, wherein the target object is located in the first layer, the second image to be processed is located in the second layer, and the first layer covers the second layer.

The fifth aspect of this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the methods described above.

As can be seen from the above technical solutions, the embodiments of this application have the following advantages:

This application provides a method for medical image processing. First, a color medical image to be processed is acquired, comprising first image data, second image data, and third image data. The first, second, and third image data correspond to color information under different attributes. Then, a difference image is generated based on the first, second, and third image data included in the medical image to be processed. This difference image is further binarized to obtain a binarized image. Finally, a foreground segmentation result corresponding to the medical image to be processed is generated based on the binarized image. Through this method, since the color information difference of grayscale pixels in different channels is small, while the color information difference of color pixels in different channels is large, a difference image is generated using the color information of different channels before binarizing the image. This effectively utilizes the color information in the image, resulting in more accurate extraction of pathological tissue regions based on the difference image, and positively impacting subsequent image analysis.

Attached Figure Description

Figure 1 is a schematic diagram of the architecture of a medical image processing system in an embodiment of this application;

Figure 2 is a schematic diagram of an embodiment of the medical image processing method in this application;

Figure 3 is a schematic diagram of an embodiment of the medical image to be processed in this application;

Figure 4 is a schematic diagram of an embodiment of the difference image in this application;

Figure 5 is a schematic diagram of an embodiment of the binarized image in this application;

Figure 6 is a schematic diagram of a foreground segmentation result in an embodiment of this application;

Figure 7 is a schematic diagram of another embodiment of the foreground segmentation result in this application;

Figure 8 is a schematic diagram of an embodiment of acquiring a medical image to be processed in this application;

Figure 9 is a flowchart illustrating a medical image processing method in an embodiment of this application.

Figure 10 is a schematic diagram of a foreground segmentation result in an embodiment of this application;

Figure 11 is a schematic diagram of an embodiment of the image processing method in this application;

Figure 12 is a schematic diagram of one embodiment of the medical image processing device in this application;

Figure 13 is a schematic diagram of an embodiment of the image processing device in this application;

Figure 14 is a schematic diagram of a server structure provided in an embodiment of this application.

Detailed Implementation

This application provides a medical image processing method, image processing method and apparatus, which generates a difference image using color information from different channels before binarizing the image, thereby effectively utilizing the color information in the image. The pathological tissue region extracted based on the difference image is more accurate and has a positive impact on subsequent image analysis.

The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “corresponding to,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

It should be understood that the embodiments of this application can be applied to scenarios involving image processing. Images, as the visual basis for human perception of the world, are an important means for humans to acquire, express, and transmit information. Image processing is a technique that analyzes images to achieve desired results. Image processing generally refers to the processing of digital images, which are large two-dimensional arrays obtained by capturing images using industrial cameras, video cameras, and scanners. The elements of this array are called pixels, and their values are called grayscale values. Image processing technology can help people understand the world more objectively and accurately. The human visual system helps humans acquire a large amount of information from the outside world, and images and graphics are the carriers of all visual information. Although the human eye has high discrimination ability and can recognize thousands of colors, in many cases, images are blurry or even invisible to the human eye. Therefore, image processing technology can make blurry or even invisible images clear. Specifically, image processing technology can include, but is not limited to, image transformation, image encoding and compression, image enhancement and restoration, image segmentation, image description, image matting, and image classification.

Specifically, the image processing method provided in this application can be applied to medical scenarios. The segmentable medical images include, but are not limited to, brain images, heart images, chest images, and cell images. Medical images may be affected by noise, field shift effects, local volume effects, and tissue motion. Due to the differences between individual organisms and the complex shapes of tissue structures, medical images typically have higher blurriness and non-uniformity compared to ordinary images. The medical images involved in this application are color images, which can be color ultrasound images or whole-slide images (WSI) images, or color digital images obtained from a microscope. Taking WSI images as an example, the side length of WSI images is typically between 10,000 and 100,000 pixels. WSI images often need to be scaled or cropped into smaller images for further processing. During image processing, it is necessary to segment the areas containing pathological tissue sections, and then perform pathological analysis based on these areas, such as quantitative analysis of cell nuclei, cell membranes, cytoplasm, and tissue microvascular analysis. Therefore, based on the characteristics of medical images, the medical image processing method of this application can acquire the medical image to be processed, and generate a difference image based on the first image data, second image data, and third image data included in the medical image to be processed. The first image data, second image data, and third image data correspond to color information under different attributes, respectively. The difference image is further binarized to obtain a binarized image. Finally, the foreground segmentation result corresponding to the medical image to be processed is generated based on the binarized image. Since the color information difference of grayscale pixels in different channels is small, while the color information difference of color pixels in different channels is large, a difference image is generated using the color information of different channels before binarizing the image. This effectively utilizes the color information in the image, resulting in more accurate extraction of pathological tissue regions based on the difference image, and positively impacting subsequent image analysis.

In another example, image processing can also be applied to remote sensing scenarios. Due to the rapid development of information technology and space technology, and the continuous improvement of satellite spatial resolution, high-resolution remote sensing images can be applied to marine monitoring, land cover monitoring, marine pollution control, and maritime rescue. High-resolution remote sensing images are characterized by rich image detail, significant geometric structures of ground features, and complex target structures. For example, in high-resolution remote sensing images, the shadows of objects along the coastline are complex, the vegetation coverage area is large, or the distinction between light and dark man-made structures is not clear. Because high-resolution remote sensing images are usually more detailed and complex than ordinary images, when it is necessary to determine the vegetation coverage area in a high-resolution remote sensing image, the vegetation can be subtracted from the high-resolution remote sensing image to determine the corresponding area. Therefore, based on the characteristics of high-resolution remote sensing images, the image processing method of this application can generate a difference image based on the first image data, second image data, and third image data included in the first image to be processed. The first image to be processed is a color image, and the first image data, second image data, and third image data included in the first image to be processed correspond to color information in different channels. Then, the generated difference image is binarized to obtain a binarized image. Based on the binarized image, a foreground segmentation result corresponding to the first image to be processed is generated. Furthermore, based on the foreground segmentation result, the target object (such as a vegetation area) is extracted from the first image to be processed. Before binarizing the image, a difference image is generated using the color information of different channels. Since the color information difference of grayscale pixels in different channels is small, while the color information difference of color pixels in different channels is large, the color information in the image can be effectively utilized. The target object extracted based on the difference image is more accurate, and the details in the high-resolution remote sensing image can be obtained more precisely, thereby improving the accuracy of high-resolution remote sensing image processing.

This application uses a medical field scenario as an example for illustration. To improve the accuracy of extracted pathological tissue regions in a medical setting and positively impact subsequent image analysis, this application proposes a medical image processing method. This method is applied to the medical image processing system shown in Figure 1. Figure 1 is a schematic diagram of the architecture of the medical image processing system in this application embodiment. As shown, the image processing system includes a server and terminal devices. The medical image processing device can be deployed on the server or on a terminal device with high computing power.

Taking a medical image processing device deployed on a server as an example, the server acquires the medical image to be processed. Then, based on the first image data, second image data, and third image data included in the medical image to be processed, the server generates a difference image. Further, the difference image is binarized to obtain a binarized image. Finally, the server generates the foreground segmentation result corresponding to the medical image to be processed based on the binarized image. The server can then perform medical image analysis based on the foreground segmentation result.

Taking a medical image processing device deployed on a terminal device as an example, the terminal device acquires the medical image to be processed. Then, based on the first image data, second image data, and third image data included in the medical image to be processed, the terminal device generates a difference image. Further, the difference image is binarized to obtain a binarized image. Finally, the terminal device generates the foreground segmentation result corresponding to the medical image to be processed based on the binarized image. The terminal device can then perform medical image analysis based on the foreground segmentation result.

In Figure 1, the server can be a single server, a server cluster consisting of multiple servers, or a cloud computing center, etc., and there are no specific limitations here. The terminal device can be a tablet computer, laptop computer, PDA, mobile phone, personal computer (PC), or voice interaction device shown in Figure 1, or it can be a monitoring device, facial recognition device, etc., and there are no limitations here.

Although only five terminal devices and one server are shown in Figure 1, it should be understood that the example in Figure 1 is only for understanding this scheme, and the specific number of terminal devices and servers should be flexibly determined based on the actual situation.

Since the embodiments of this application are applied to the field of artificial intelligence, before introducing the model training method provided in the embodiments of this application, some basic concepts in the field of artificial intelligence will be introduced first. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain optimal results. In other words, artificial intelligence is a comprehensive technology of computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. Artificial intelligence studies the design principles and implementation methods of various intelligent machines, enabling machines to have the functions of perception, reasoning, and decision-making. Artificial intelligence technology is a comprehensive discipline involving a wide range of fields, including both hardware and software technologies. Basic artificial intelligence technologies generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operating/interactive systems, mechatronics, and other technologies. Artificial intelligence software technologies mainly include computer vision technology, speech processing technology, natural language processing technology, and machine learning/deep learning. Machine learning (ML) is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory, among others. It specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and instructional learning.

With the research and advancement of artificial intelligence (AI) technology, AI research has expanded into various directions. Computer vision (CV) is one such research area within AI, studying how to enable machines to "see." More specifically, it refers to machine vision, which uses cameras and computers to replace human eyes in tasks such as target recognition, tracking, and measurement, and further performs image processing to create images more suitable for human observation or transmission to instruments. As a scientific discipline, computer vision researches related theories and technologies, attempting to establish AI systems capable of extracting information from images or multidimensional data. Computer vision technologies typically include image processing, image recognition, image semantic understanding, image retrieval, optical character recognition (OCR), video processing, video semantic understanding, video content/behavior recognition, 3D object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping (SLAM), and common biometric recognition technologies such as facial recognition and fingerprint recognition.

The solutions provided in this application relate to image processing technology using artificial intelligence. Based on the above description, the method for medical image processing in this application will be described below. Please refer to Figure 2, which is a schematic diagram of an embodiment of the medical image processing method in this application. As shown in the figure, an embodiment of the medical image processing method in this application includes:

101. Acquire a medical image to be processed, wherein the medical image to be processed is a color image, and the medical image to be processed includes first image data, second image data and third image data, and the first image data, second image data and third image data correspond to color information under different channels respectively;

In this embodiment, the medical image processing device can acquire a color medical image to be processed. This medical image may include first image data, second image data, and third image data, each corresponding to color information in a different channel. The medical image to be processed can be a medical image received by the medical image processing device via a wired network, or a medical image stored within the medical image processing device itself.

Specifically, the medical image to be processed can be a region cropped from a WSI image. This WSI image can be obtained by scanning a slide under a microscope. Since the slide refers to a glass slide prepared after staining with hematoxylin or other methods, the WSI image obtained after scanning the slide under a microscope is a color image. The color mode of the color image includes, but is not limited to, the red-green-blue (RGB) color mode, the luminance-bandwidth-chrominance (YUV) color mode, and the hue-saturation-value (HSV) color mode. Color information can be represented as pixel values in different channels, such as the pixel values of the R channel, G channel, and B channel.

WSI images come in formats including, but not limited to, SVS and NDPI. Since WSI images are typically tens of thousands of pixels in length and width, their large size necessitates significant memory usage for direct processing. Therefore, WSI image cropping is necessary. Python's OpenSlide tool can be used to read WSI images. OpenSlide can convert file formats and store a cropped region as a 12x12 resolution image. In practice, this includes, but is not limited to, storing it at resolutions of 15x15 and 50x50. Multiple images can reside in the same WSI image file. In practical applications, the image with the highest resolution in the WSI image file is selected as the image to be processed. Furthermore, this embodiment can crop the medical image to be processed from a reduced WSI image, and the WSI image can be reduced by any factor, such as 20x or 10x. The length and width of the reduced WSI image are within the range of several thousand pixels. It should be understood that since the reduction factor is manually defined, the specific reduction factor should be flexibly determined based on the actual situation.

For ease of understanding, please refer to Figure 3. Figure 3 is a schematic diagram of an embodiment of the medical image to be processed in this application. As shown in the figure, the medical image to be processed includes a case tissue area, and there are no other grayscale or pure white backgrounds interfering with the medical image to be processed. To further understand this embodiment, we will use RGB as an example for illustration. Since the first image data, second image data, and third image data included in the medical image to be processed correspond to color information under different channels, if the RGB corresponding to the color image is (200, 100, 60), then the first image data can be the pixel value 200 corresponding to the R channel, the second image data can be the pixel value 100 corresponding to the G channel, and the third image data can be the pixel value 60 corresponding to the B channel. If the RGB corresponding to the color image is (100, 80, 40), then the first image data can be the pixel value 100 corresponding to the R channel, the second image data can be the pixel value 800 corresponding to the G channel, and the third image data can be the pixel value 40 corresponding to the B channel.

It should be noted that for HSV or YUV images, you can first convert the HSV or YUV image to an RGB image before performing subsequent processing.

It should be understood that in practical applications, the specific color information corresponding to the first, second, and third image data should be flexibly determined based on the actual situation. Furthermore, the medical image processing device can be deployed on a server or on a terminal device with high computing power; this embodiment uses the deployment of the medical image processing device on a server as an example.

102. Generate a difference image based on the first image data, the second image data, and the third image data included in the medical image to be processed;

In this embodiment, the medical image processing device can generate a difference image based on the first image data, the second image data, and the third image data. Specifically, the difference image is a grayscale image. For ease of understanding, please refer to Figure 4, which is a schematic diagram of an embodiment of the difference image in this application. As shown in the figure, based on the medical image to be processed, including the first image data, the second image data, and the third image data shown in Figure 4 (A), the difference image shown in Figure 4 (B) can be generated. This difference image can be an image including pathological tissue areas as shown in the figure. Since color information corresponding to different channels is used to distinguish pixel values, taking the image color mode of the medical image to be processed as RGB as an example, if the medical image to be processed is grayscale, the RGB values are relatively similar; if the medical image to be processed is color, the difference between RGB values is large, and there is a large color difference in the areas where pathological tissue exists.

103. Perform binarization on the difference image to obtain a binarized image;

In this embodiment, the medical image processing device can binarize the difference image generated in step 102 to obtain a binarized image. For ease of understanding, please refer to Figure 5, which is a schematic diagram of an embodiment of the binarized image in this application. As shown in Figure 5(A), since the difference image is a grayscale image, grayscale-based segmentation can be used. Specifically, in this embodiment, an adaptive binarization method is used for foreground segmentation, that is, the difference image is binarized to obtain the binarized image shown in Figure 5(B). Furthermore, in this binarized image, white represents the foreground region including the pathological tissue region, while black represents the background region excluding the pathological tissue region.

104. Generate the foreground segmentation result corresponding to the medical image to be processed based on the binarized image.

In this embodiment, the medical image processing device can generate a foreground segmentation result corresponding to the medical image to be processed based on the binarized image obtained in step 103. For ease of understanding, please refer to Figure 6, which is a schematic diagram of an embodiment of the foreground segmentation result in this application. As shown in the figure, based on the binarized image shown in Figure 6(A), since white represents the foreground region including the pathological tissue region and black represents the background region excluding the pathological tissue region, the foreground segmentation result corresponding to the medical image to be processed shown in Figure 6(B) can be generated based on the binarized image.

In this embodiment of the application, a method for medical image processing is provided. In this method, since the color information difference of grayscale pixels in different channels is small, while the color information difference of color pixels in different channels is large, before binarizing the image, a difference image is generated using the color information of different channels. This effectively utilizes the color information in the image, and the pathological tissue region extracted based on the difference image is more accurate, which has a positive impact on subsequent image analysis.

Optionally, based on the embodiment corresponding to Figure 2 above, in an optional embodiment of the medical image processing method provided in this application, generating a difference image according to the first image data, second image data, and third image data included in the medical image to be processed may include:

A maximum value image is generated based on the first image data, the second image data, and the third image data included in the medical image to be processed.

A minimum value image is generated based on the first image data, the second image data, and the third image data included in the medical image to be processed.

Generate a difference image based on the maximum and minimum value images.

In this embodiment, after acquiring the medical image to be processed, the medical image processing device can generate a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed. Then, it can generate a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed. Finally, a difference image can be generated based on the maximum value image and the minimum value image.

Specifically, taking the RGB color mode of the medical image to be processed as an example, since the first image data, the second image data, and the third image data of the medical image to be processed correspond to color information in different channels, the color information is represented by the pixel values corresponding to the R channel, G channel, and B channel. The maximum value in the R channel, G channel, and B channel is determined, and the maximum value image is determined by the maximum value. Similarly, the minimum value in the R channel, G channel, and B channel can be determined, and the minimum value image can be determined by the minimum value. Then, each pixel in the maximum value image is subtracted from each pixel in the corresponding position in the minimum value image to obtain the difference image.

In this embodiment of the application, a method for generating a difference image is provided. By means of the above method, a maximum value image and a minimum value image are generated based on the first image data, the second image data, and the third image data. Since the color information corresponding to different image data is different, the maximum value image and the minimum value image determined based on different image data include color information of the medical image to be processed with high accuracy, thereby improving the accuracy of the difference image generation.

Optionally, based on the embodiment corresponding to Figure 2 above, in another optional embodiment of the medical image processing method provided in this application, generating a maximum value image based on the first image data, second image data, and third image data included in the medical image to be processed may include:

Based on the first image data, second image data, and third image data included in the medical image to be processed, the maximum pixel value corresponding to the target pixel is determined from the first pixel value, the second pixel value, and the third pixel value. The first pixel value is the pixel value corresponding to the first pixel position in the first image data, the second pixel value is the pixel value corresponding to the second pixel position in the second image data, the third pixel value is the pixel value corresponding to the third pixel position in the third image data, and the target pixel is the pixel value corresponding to the fourth pixel position in the maximum value image. The first pixel position, the second pixel position, the third pixel position, and the fourth pixel position all correspond to the position of the same pixel in the medical image to be processed.

Generating a minimum value image based on the first image data, second image data, and third image data included in the medical image to be processed may include:

Based on the first image data, second image data, and third image data included in the medical image to be processed, determine the minimum pixel value corresponding to the target pixel from the first pixel value, second pixel value, and third pixel value;

Generating a difference image based on the maximum and minimum value images can include:

Subtract the maximum pixel value corresponding to the target pixel in the maximum value image from the minimum pixel value corresponding to the target pixel in the maximum value image to obtain the difference pixel value corresponding to the target pixel in the difference image.

In this embodiment, the medical image processing device can determine the maximum pixel value corresponding to the target pixel based on the first image data, second image data, and third image data included in the medical image to be processed. Secondly, it can also determine the minimum pixel value corresponding to the target pixel based on the first image data, second image data, and third image data included in the medical image to be processed. Then, based on the determined maximum and minimum pixel values, the maximum pixel value corresponding to the target pixel in the maximum value image is subtracted from the minimum pixel value corresponding to the target pixel in the maximum value image to obtain the difference pixel value corresponding to the target pixel in the difference image.

For ease of understanding, let's take an RGB color mode medical image as an example. For the medical image containing first, second, and third image data, each pixel has corresponding image data in the R, G, and B channels. For example, the pixel's image data in the R channel is the first pixel value, in the G channel is the second pixel value, and in the B channel is the third pixel value. Based on these values, the maximum pixel value in the R, G, and B channels can be determined. Similarly, the minimum pixel value in the R, G, and B channels can be determined based on these values.

Furthermore, the maximum and minimum pixel values at pixel position (x, y) can be calculated using the following formula:

Imax(x,y)=Max[Ir(x,y),Ig(x,y),Ib(x,y)];

Imin(x,y)=Min[Ir(x,y),Ig(x,y),Ib(x,y)];

Where Imax(x, y) represents the maximum pixel value, Imin(x, y) represents the minimum pixel value, Ir(x, y) represents the first pixel value, Ig(x, y) represents the second pixel value, and Ib(x, y) represents the third pixel value.

This embodiment uses color information corresponding to different channels to distinguish pixel values. Again, taking the RGB color mode of the medical image to be processed as an example, if the medical image is grayscale, the RGB values are relatively similar; if the medical image is color, the differences between RGB values are large. Areas containing pathological tissue have color, and the values with color are the pixel values required in this embodiment. It should be understood that the aforementioned formula is only used as an example for pixels corresponding to a two-dimensional image. In practical applications, the aforementioned formula is also applicable to calculating the maximum and minimum pixel values of multi-dimensional images, such as three-dimensional (3D) images and four-dimensional (4D) images.

To further understand this embodiment, we will use an example where the target pixel position is (x1, y1) and the color mode of the medical image to be processed is RGB. The first pixel value Ir(x1, y1) of the target pixel position (x1, y1) is 100, the second pixel value Ig(x1, y1) of the target pixel position (x1, y1) is 200, and the third pixel value Ib(x1, y1) of the target pixel position (x1, y1) is 150. According to the aforementioned formula, the maximum pixel value Imax(x1, y1) of the target pixel position (x1, y1) is the pixel value 200 corresponding to the second pixel value Ig(x1, y1), and the minimum pixel value Imin(x1, y1) of the target pixel position (x1, y1) is the pixel value 100 corresponding to the first pixel value Ir(x1, y1). Secondly, taking the target pixel position (x2, y2) and the image color mode of the medical image to be processed as another example, the first pixel value Ir(x2, y2) of the target pixel position (x2, y2) is 30, the second pixel value Ig(x2, y2) of the target pixel position (x2, y2) is 80, and the third pixel value Ib(x2, y2) of the target pixel position (x2, y2) is 120. According to the aforementioned formula, the maximum pixel value Imax(x2, y2) of the target pixel position (x2, y2) is the pixel value of 120 corresponding to the third pixel value Ib(x2, y2), and the minimum pixel value Imin(x2, y2) of the target pixel position (x2, y2) is the pixel value of 30 corresponding to the first pixel value Ir(x2, y2). Furthermore, taking a medical image to be processed as an example, with the image color mode being RGB and the image being a 3D image, and the target pixel position being (x3, y3, z3), the first pixel value Ir(x3, y3, z3) at the target pixel position (x3, y3, z3) is 200, and the second pixel value Ig(x3, y3, z3) at the target pixel position (x3, y3, z3) is 10. 3) The third pixel value Ib(x3, y3, z3) is 60. According to the aforementioned formula, the maximum pixel value Imax(x3, y3, z3) at the target pixel position (x3, y3, z3) is the pixel value 200 corresponding to the first pixel value Ir(x3, y3, z3), and the minimum pixel value Imin(x3, y3, z3) at the target pixel position (x3, y3, z3) is the pixel value 10 corresponding to the second pixel value Ig(x3, y3, z3).

Furthermore, after obtaining the maximum and minimum pixel values corresponding to the target pixel location, the maximum and minimum pixel values can be subtracted to obtain the difference pixel value corresponding to the target pixel location in the difference image. Specifically, the difference pixel value can be calculated using the following formula based on the maximum pixel value Imax(x, y) and the minimum pixel value Imin(x, y), assuming that the medical image to be processed contains 10,000 pixels:

Idiff(x,y)=Imax(x,y)-Imin(x,y);

Where Imax(x, y) represents the maximum pixel value, Imin(x, y) represents the minimum pixel value, and Idiff(x, y) represents the difference pixel value at position (x, y).

For ease of understanding, we will use the target pixel position (x1, y1) and the color mode of the medical image to be processed as an example. The maximum pixel value Imax(x1, y1) at the target pixel position (x1, y1) is 200, and the minimum pixel value Imin(x1, y1) at the target pixel position (x1, y1) is 100. Subtracting the maximum pixel value Imax(x1, y1) from the minimum pixel value Imin(x1, y1) will give us the difference pixel value of 100 at the target pixel position (x1, y1). Secondly, taking the target pixel position (x2, y2) and the image color mode of the medical image to be processed as another example, the maximum pixel value Imax(x2, y2) at the target pixel position (x2, y2) is 120, and the minimum pixel value Imin(x2, y2) at the target pixel position (x2, y2) is 30. Subtracting the maximum pixel value Imax(x2, y2) from the minimum pixel value Imin(x2, y2) gives the difference pixel value corresponding to the target pixel position (x2, y2) as 90.

Optionally, another example is provided, where the medical image to be processed is in RGB color mode and is a 3D image, with the target pixel position being (x3, y3, z3). Based on the above formula, the following formula can be derived:

Imax(x,y,z)=Max[Ir(x,y,z),Ig(x,y,z),Ib(x,y,z)];

Imin(x,y,z)=Min[Ir(x,y,z),Ig(x,y,z),Ib(x,y,z)];

Idiff(x,y,z)=Imax(x,y,z)-Imin(x,y,z);

Assuming the maximum pixel value Imax(x3, y3, z3) at the target pixel position (x3, y3, z3) is 200, and the minimum pixel value Imin(x3, y3, z3) at the target pixel position (x3, y3, z3) is 10, subtracting the maximum pixel value Imax(x3, y3, z3) from the minimum pixel value Imin(x3, y3, z3) yields a difference pixel value of 190 corresponding to the target pixel position (x3, y3, z3).

Specifically, when the difference pixel value of the medical image to be processed is small, it indicates that the first, second, and third pixel values of the medical image to be processed are relatively similar, which means that the medical image to be processed is similar to a grayscale image. When the difference pixel value of the medical image to be processed is large, it indicates that the first, second, and third pixel values of the medical image to be processed are significantly different, which means that the medical image to be processed is similar to a color image. Since images containing pathological tissue areas are often colored images, the difference pixel value can be used to make a preliminary judgment on whether the medical image to be processed includes pathological tissue areas.

In this embodiment of the application, a method for generating a maximum value image is provided. In the above manner, the maximum pixel value and the minimum pixel value are determined by the pixel values of the target pixel points corresponding to the first image data, the second image data, and the third image data. The maximum pixel value and the minimum pixel value reflect the color information of the medical image to be processed to different degrees. The difference pixel value is obtained by subtracting the maximum pixel value and the minimum pixel value, so that the difference pixel value can accurately reflect the color information of the medical image to be processed, thereby improving the accuracy of the difference image generation.

Optionally, based on the embodiment corresponding to Figure 2 above, in another optional embodiment of the medical image processing method provided in this application, generating a difference image according to the first image data, second image data, and third image data included in the medical image to be processed may include:

Based on the first image data, second image data, and third image data included in the medical image to be processed, a difference image to be processed is generated.

Gaussian blur is applied to the difference image to obtain the difference image.

In this embodiment, the medical image processing device can generate a difference image to be processed based on the first image data, the second image data, and the third image data included in the medical image to be processed, and then perform Gaussian blur processing on the difference image to be processed to obtain the difference image.

Specifically, blurring can be understood as taking the average of the surrounding pixels for each pixel in the difference image to be processed. When a pixel takes the average of its surrounding pixels, the pixel value tends to be smoothed out. However, in the difference image to be processed, this is equivalent to producing a blurring effect, causing the pixel to lose detail. In the difference image to be processed, the pixels are continuous; therefore, pixels closer together are more closely related, and pixels farther apart are less related. Therefore, this embodiment uses Gaussian blur as the blurring algorithm. Gaussian blur applies a normal distribution (Gaussian distribution) to the difference image to be processed, making the weighted average between pixels more reasonable, with closer pixels having a higher weight and farther pixels having a lower weight.

Furthermore, taking pixel (x, y) as an example, for pixel (x, y), this pixel is a two-dimensional pixel, and the two-dimensional Gaussian function can be calculated using the following formula:

Where (x, y) represents a pixel, G(x, y) represents the two-dimensional Gaussian function of the pixel, and σ represents the standard deviation of the normal distribution.

For ease of understanding, let's take the pixel (0, 0) as an example. Then, the eight pixels surrounding the pixel (0, 0) can be (-1, 1), (0, 1), (1, 1), (-1, 0), (1, 0), (-1, -1), (0, -1), and (1, -1). To further calculate the weight matrix, we need to set the value of σ. Assuming σ = 1.5, we can obtain a weight matrix with a blur radius of 1. For example, in the weight matrix, the weight corresponding to pixel (0, 0) is 0.0707, the weight corresponding to pixel (-1, 1) is 0.0453, the weight corresponding to pixel (0, 1) is 0.0566, the weight corresponding to pixel (1, 1) is 0.0453, the weight corresponding to pixel (-1, 0) is 0.0566, the weight corresponding to pixel (1, 0) is 0.0566, the weight corresponding to pixel (-1, -1) is 0.0453, the weight corresponding to pixel (0, -1) is 0.0566, and the weight corresponding to pixel (1, -1) is 0.0453. The sum of the weights of the 8 pixels surrounding pixel (0, 0) is approximately 0.479. If we only calculate the weighted average of these 9 points, we need to make their weight sum equal to 1, that is... Normalizing the sum of weights involves dividing each of the nine values in the weight matrix by the sum of weights, 0.479, to obtain the normalized weight matrix. Specifically, the normalized weight for pixel (0,0) is 0.147, for (-1,1) it's 0.0947, for (0,1) it's 0.0118, for (1,1) it's 0.0947, for (-1,0) it's 0.0118, for (1,0) it's 0.0118, for (-1,-1) it's 0.0947, for (0,-1) it's 0.0118, and for (1,-1) it's 0.0947. Since using a weight matrix with a total weight greater than 1 will make the difference image brighter, while using a weight matrix with a total weight less than 1 will make the difference image darker, the normalized weight matrix can make the pathological tissue area presented by the difference image more accurate.

Furthermore, once the normalized weight matrix is obtained, Gaussian blur can be calculated for the pixel. For example, when the gray value is between 0 and 255, in the weight matrix, the gray value corresponding to pixel (0, 0) is 25, the gray value corresponding to pixel (-1, 1) is 14, the gray value corresponding to pixel (0, 1) is 15, the gray value corresponding to pixel (1, 1) is 16, the gray value corresponding to pixel (-1, 0) is 24, the gray value corresponding to pixel (1, 0) is 26, the gray value corresponding to pixel (-1, -1) is 34, the gray value corresponding to pixel (0, -1) is 35, and the gray value corresponding to pixel (1, -1) is 36. Multiplying the grayscale value of each pixel by its corresponding weight yields nine values: 3.69 for pixel (0,0), 1.32 for pixel (-1,1), 1.77 for pixel (0,1), 1.51 for pixel (1,1), 2.83 for pixel (-1,0), 3.07 for pixel (1,0), 3.22 for pixel (-1,-1), 4.14 for pixel (0,-1), and 3.41 for pixel (1,-1). Adding these nine values together gives the Gaussian blur value for pixel (0,0).

Furthermore, by repeating the steps similar to those for pixel (0,0) for all pixels in the difference image to be processed, the difference image after Gaussian blur processing can be obtained.

In this embodiment of the application, another method for generating a difference image is provided. In this method, the difference image to be processed is subjected to Gaussian blur processing. Since Gaussian blur processing can improve segmentation robustness, the resulting difference image has better segmentation robustness, thereby improving the stability of the difference image.

Optionally, based on the embodiment corresponding to Figure 2 above, in another optional embodiment of the medical image processing method provided in this application, binarizing the difference image to obtain a binarized image may include:

Determine the segmentation threshold based on the difference image;

If the pixel value corresponding to a pixel in the difference image is greater than or equal to the segmentation threshold, then the pixel is determined as the foreground pixel of the binarized image.

If the pixel value corresponding to a pixel in the difference image is less than the segmentation threshold, then the pixel is determined as the background pixel of the binarized image.

In this embodiment, the medical image processing device can determine the segmentation threshold based on the difference image. When the pixel value corresponding to a pixel in the difference image is greater than or equal to the segmentation threshold, the pixel is determined as the foreground pixel of the binarized image. When the pixel value corresponding to a pixel in the difference image is less than the segmentation threshold, the pixel is determined as the background pixel of the binarized image.

Specifically, by setting a segmentation threshold, binarizing the difference image can transform a grayscale image into a binary image with values of 0 or 1. In other words, binarizing the difference image can be achieved by setting a segmentation threshold, transforming it into a binary image where the foreground and background are represented by only two values (0 or 1), where the foreground value is 1 and the background value is 0. In practical applications, 0 corresponds to all RGB values being 0, and 1 corresponds to all RGB values being 255. After binarization, the resulting binary image can be further processed because its geometric properties depend only on the positions of 0 and 1, no longer involving the pixel's grayscale value. This simplifies the processing of the binary image and improves image processing efficiency. Methods for determining the segmentation threshold can be divided into global thresholding and local thresholding. Global thresholding divides the entire difference image using a single threshold. However, the grayscale depth of different difference images varies, and the brightness distribution of different parts of the same difference image can also be different. Therefore, in this embodiment, we use a dynamic threshold binarization method to determine the segmentation threshold.

After determining the segmentation threshold based on the difference image, the pixel values corresponding to pixels in the difference image are compared with the segmentation threshold. If the pixel value is greater than or equal to the segmentation threshold, the pixel is designated as a foreground pixel in the binarized image. If the pixel value is less than the segmentation threshold, the pixel is designated as a background pixel in the binarized image. For example, if the pixel value of pixel A is greater than the segmentation threshold, pixel A is designated as a foreground pixel in the binarized image (i.e., a pixel value of 1), meaning pixel A is in the foreground region and will appear white in RGB mode. Conversely, if the pixel value of pixel B is less than the segmentation threshold, pixel B is designated as a background pixel in the binarized image (i.e., a pixel value of 0), meaning pixel B is in the background region and will appear black in RGB mode.

In this embodiment of the application, a method for obtaining a binarized image is provided. By means of the above method, a binarized image is generated according to the binarization process. Since the geometric properties of the binarized image do not involve the gray values of the pixels, the subsequent processing of the binarized image can be simplified, thereby improving the efficiency of generating foreground segmentation results.

Optionally, based on the embodiment corresponding to Figure 2 above, in another optional embodiment of the medical image processing method provided in this application, determining the segmentation threshold based on the difference image may include:

Obtain N pixel values corresponding to N pixels from the difference image, where there is a one-to-one correspondence between pixel values and pixel points, and N is an integer greater than 1.

Determine the pixel value to be processed from N pixel values, where the pixel value to be processed is the maximum value among the N pixel values;

The segmentation threshold is calculated based on the pixel value to be processed and the ratio threshold.

In this embodiment, the medical image processing device can obtain N pixel values corresponding to N pixels based on the difference image, and the pixel value has a one-to-one correspondence with the pixel. Then, it determines the pixel value to be processed from the N pixel values. The pixel value to be processed is the maximum value among the N pixel values. Finally, it calculates the segmentation threshold based on the pixel value to be processed and the ratio threshold, where N is an integer greater than 1.

Specifically, in this embodiment, the segmentation threshold is determined based on the difference image. Since the difference image can be generated by subtracting the maximum and minimum value images from the medical image to be processed, and there is a one-to-one correspondence between pixel values and pixels in the difference image, multiple pixel values can be obtained from the difference image. Then, the maximum value among these multiple pixel values is determined as the pixel value to be processed. The segmentation threshold is then calculated based on the pixel value to be processed and a proportional threshold. For ease of understanding, this embodiment uses a proportional threshold of 10% as an example. For instance, if the width and height of the WSI image after reduction are within a few thousand pixels, assuming the reduced image includes 100*100 pixels, the largest value needs to be found among the 10,000 pixel values. For example, if the maximum value is 150, then this maximum value of 150 can be determined as the pixel value to be processed. Then, by multiplying the pixel value to be processed (150) by the proportional threshold of 10%, the segmentation threshold of 15 is obtained. It should be understood that in practical applications, the proportional threshold can also be other percentage values, and the specific proportional threshold should be flexibly determined based on the actual situation.

In this embodiment of the application, another method for obtaining the segmentation threshold is provided. In the above manner, the segmentation threshold can be determined by the pixel value to be processed by the maximum pixel value and the ratio threshold. Since the gray depth of the difference image is different and the brightness distribution of different regions can also be different, the segmentation threshold can be flexibly determined by adjusting the ratio threshold, thereby improving the accuracy and flexibility of the threshold and thus improving the accuracy of binarized image generation.

Optionally, based on the embodiment corresponding to Figure 2 above, in another optional embodiment of the medical image processing method provided in this application, generating the foreground segmentation result corresponding to the medical image to be processed based on the binarized image may include:

A flooding algorithm is used to detect background regions in a binarized image, where the background region includes multiple background pixels.

Based on the binarized image and the background region in the binarized image, obtain the background pixels within the foreground region of the binarized image, wherein the foreground region includes multiple foreground pixels;

By replacing background pixels in the foreground region of a binarized image with foreground pixels, a hole-filling image is obtained.

Median filtering is applied to the hole-filling image to obtain the foreground segmentation result corresponding to the medical image to be processed.

In this embodiment, the medical image processing device can use a flooding algorithm to detect the background region in the binarized image. The background region may include multiple background pixels. Then, based on the binarized image and the background region in the binarized image, the background pixels in the foreground region of the binarized image are obtained. The foreground region may include multiple foreground pixels. Then, the background pixels in the foreground region of the binarized image are changed into foreground pixels to obtain a hole-filling image. Finally, the hole-filling image is subjected to median filtering to obtain the foreground segmentation result corresponding to the medical image to be processed.

Specifically, after binarizing the difference image, the resulting binarized image may contain black holes in the foreground region. These black holes need to be detected. For clarity, please refer to Figure 7, which is a schematic diagram of another embodiment of the foreground segmentation result in this application. As shown in Figure 7(A), the binarized image includes multiple background pixels in the white foreground region. The black dots enclosed in regions A1 to A5 are all composed of background pixels. By changing the black dots enclosed in regions A1 to A5 from background pixels to foreground pixels, and the white dots enclosed in regions A6 and A7 from foreground pixels, the hole-filling image shown in Figure 7(B) can be obtained.

Furthermore, median filtering is then applied to the hole-filling image shown in Figure 7(B), and morphological processing can be further performed to obtain the foreground segmentation result corresponding to the medical image to be processed, as shown in Figure 7(C). The filtering process aims to suppress noise in the medical image to be processed while preserving as much detail as possible in the hole-filling image. Filtering improves the effectiveness and reliability of subsequent foreground segmentation processing and analysis. Eliminating noise components in the hole-filling image is the filtering operation. The energy of the hole-filling image is mostly concentrated in the low and mid-frequency bands of the amplitude spectrum, while in higher frequency bands, the information in the hole-filling image is often affected by noise. Therefore, filtering can be applied to the hole-filling image to meet the requirements of image processing and eliminate noise introduced during image digitization. Median filtering is a typical nonlinear filtering technique based on ranking statistics theory, which effectively suppresses noise. Median filtering uses the median gray value of a pixel's neighborhood to replace the pixel's gray value, making the surrounding pixel values closer to the true values, thereby eliminating isolated noise points.

In this embodiment of the application, a method for generating foreground segmentation results is provided. By changing the background pixels in the foreground region to foreground pixels in the above manner, the resulting hole-filling image has good reliability. Secondly, by median filtering, the foreground segmentation result corresponding to the medical image to be processed can be clear and have good visual effect without damaging the contour and edge features of the image.

Optionally, based on the embodiment corresponding to Figure 2 above, in another optional embodiment of the medical image processing method provided in this application, performing median filtering on the hole-filling image to obtain the foreground segmentation result corresponding to the medical image to be processed may include:

The hole-filling image is subjected to median filtering to obtain a filtered image, which includes the foreground region to be processed.

Obtain the boundary line of the foreground region to be processed, where the boundary line consists of M pixels, M being an integer greater than 1;

For each of the M pixels on the boundary line, extend outward by K pixels to obtain the foreground segmentation result, where K is an integer greater than or equal to 1.

In this embodiment, the medical image processing device performs median filtering on the hole-filling image to obtain a filtered image. This filtered image may include a foreground region to be processed. The boundary line of the foreground region is obtained, and this boundary line includes M pixels. Then, for each of the M pixels on the boundary line, K pixels are extended outwards to obtain the foreground segmentation result, where M is an integer greater than 1 and K is an integer greater than or equal to 1. Specifically, median filtering can replace the gray value of a pixel with the median of its neighborhood gray values, making the surrounding pixel values closer to the true values, thereby eliminating isolated noise points. Through median filtering, while removing impulse noise and salt-and-pepper noise, a filtered image that preserves the edge details of the image is obtained.

Furthermore, a flood fill algorithm is used to fill connected, similarly colored regions with different colors. The basic principle of the flood fill algorithm is to start from a pixel and expand the coloring outwards to surrounding pixels until the boundary of the graphic. The flood fill algorithm requires three parameters: a start node, a target color, and a replacement color. The flood fill algorithm connects all nodes to the start node along the path of the target color and changes them to the replacement color. It should be understood that in practical applications, flood fill algorithms can be constructed in various ways, but many methods explicitly or implicitly use queue or stack data structures. Examples include four-neighborhood flood fill, eight-neighborhood flood fill, scanline fill, and large-scale behavior. The traditional four-neighbor flooding algorithm colors a pixel (x, y) and then colors its four surrounding pixels (up, down, left, and right). However, this recursive approach is memory-intensive, and can lead to overflow if the area to be colored is very large. Therefore, a non-recursive four-neighbor flooding algorithm can be used. The eight-neighbor flooding algorithm, on the other hand, colors all four sides of a pixel: up, down, left, right, top-left, bottom-left, top-right, and bottom-right. Line drawing algorithms can be accelerated by using fill lines. Pixels along a line can be colored first, and then the coloring can be expanded upwards and downwards until all pixels are colored. Large-scale operations are data-centric or process-centric.

Because the boundary lines of the hole-filling image are irregular, this embodiment uses a line-drawing algorithm. Taking the boundary line of the foreground region to be processed, which includes 1000 pixels, as an example, morphological processing is used to extend each of the 1000 pixels outward by K pixels. Assuming K is 2, this adds 2000 pixels outside the original 1000 pixels as the foreground segmentation region, thus obtaining the foreground segmentation result. It should be understood that in practical applications, the specific M pixels and K pixels should be flexibly determined based on the actual situation.

This application provides another method for generating foreground segmentation results. Through median filtering, the filtered image is clear and has good visual quality without damaging the image's contours and edges. Furthermore, a flooding algorithm is used to perform morphological processing on the filtered image, improving the accuracy and uniformity of the foreground segmentation results.

Optionally, based on the embodiment corresponding to Figure 2 above, in another optional embodiment of the medical image processing method provided in this application, acquiring the medical image to be processed may include:

Acquire raw medical images;

A sliding window method is used to extract medical sub-images from the original medical image;

If a pathological tissue region is detected in a medical sub-image, it is identified as a medical image to be processed.

If a medical sub-image is found to not include a pathological tissue region, the medical sub-image is identified as a background image and removed.

In this embodiment, the medical image processing device first acquires the original medical image, and then uses a sliding window to extract medical sub-images from the original medical image. When a medical sub-image is detected to contain a pathological tissue region, it is determined to be a medical image to be processed. When a medical sub-image is detected not to contain a pathological tissue region, it is determined to be a background image and removed. Specifically, the original medical image can be an image received by the medical image processing device through a wired network, or it can be an image stored by the medical image processing device itself.

For ease of understanding, please refer to Figure 8. Figure 8 is a schematic diagram of an embodiment of obtaining a medical image to be processed in this application. As shown in the figure, (A) in Figure 8 shows the original medical image. A sliding window is used to extract medical sub-images from the original medical image. The areas enclosed by B1 to B3 are the medical sub-images extracted from the original medical image. Thus, B1 can be obtained as the medical sub-image shown in Figure 8 (B), B2 can be obtained as the medical sub-image shown in Figure 8 (C), and B3 can be obtained as the medical sub-image shown in Figure 8 (D). It can be seen that the medical sub-images shown in Figure 8 (B) and (C) include pathological tissue areas. Therefore, the medical sub-images shown in Figure 8 (B) and (C) can be identified as the medical images to be processed. The medical sub-image shown in Figure 8 (D) does not include pathological tissue areas. Therefore, the medical sub-image shown in Figure 8 (D) can be identified as the background image and the background image is removed.

This application provides a method for acquiring a medical image to be processed. By detecting whether a medical sub-image includes a pathological tissue region, the medical image to be processed is determined. This ensures that the medical image containing a pathological tissue region can be segmented into a foreground segmentation result through the aforementioned steps. Furthermore, the foreground segmentation result includes the pathological tissue region, facilitating subsequent processing and analysis of this pathological tissue region. Secondly, the medical sub-image containing the pathological tissue region is designated as a background image and removed to reduce resource consumption.

Optionally, based on the embodiment corresponding to Figure 2 above, in another optional embodiment of the medical image processing method provided in this application, after generating the foreground segmentation result corresponding to the medical image to be processed based on the binarized image, the medical image processing method may further include:

The target positive sample image is generated based on the foreground segmentation result. The target positive sample image belongs to a positive sample image in the positive sample set, and each positive sample image contains a pathological tissue region.

Obtain a negative sample set, wherein the negative sample set includes at least one negative sample image, and each negative sample image does not contain a pathological tissue region;

The image segmentation model is trained based on the positive and negative sample sets.

In this embodiment, after generating the foreground segmentation result corresponding to the medical image to be processed based on the binarized image, the medical image processing device can also generate a target positive sample image based on the foreground segmentation result. This target positive sample image belongs to a positive sample image in a set of positive samples, and each positive sample image contains a pathological tissue region. Simultaneously, a negative sample set can be obtained, which includes at least one negative sample image, and each negative sample image does not contain a pathological tissue region. Finally, the image segmentation model can be trained based on the obtained positive and negative sample sets. This image segmentation model can segment the corresponding pathological tissue region based on a color medical image.

In this embodiment of the application, a method for training an image segmentation model is provided. By using the above method, the image segmentation model is trained with a set of positive sample images containing pathological tissue regions and a set of negative sample images not containing pathological tissue regions, thereby improving the accuracy and reliability of the image segmentation model and thus improving the efficiency and accuracy of image processing.

Specifically, the embodiments of this application can improve the accuracy of extracted pathological tissue regions and have a positive impact on subsequent image analysis. For easier understanding of the embodiments of this application, please refer to Figure 9, which is a flowchart of a medical image processing method in the embodiments of this application. Specifically:

In step S1, the original medical image is acquired;

In step S2, the medical image to be processed is obtained based on the original medical image;

In step S3, a difference image is generated based on the medical image to be processed;

In step S4, the difference image is binarized to obtain a binarized image;

In step S5, a hole-filling image is obtained based on the binarized image;

In step S6, the hole-filling image is subjected to median filtering to obtain the foreground segmentation result.

In step S1, the original medical image shown in Figure 9(A) can be obtained. Then, in step S2, a sliding window is used to extract a medical sub-image from the original medical image shown in Figure 9(A). When a pathological tissue region is detected in the medical sub-image, it is determined to be a medical image to be processed, thereby obtaining the medical image to be processed shown in Figure 9(B). Further, in step S3, the maximum and minimum pixel values corresponding to the target pixel can be determined from the first pixel value, the second pixel value, and the third pixel value according to the first image data, the second image data, and the third image data included in the medical image to be processed, thereby generating a maximum value image and a minimum value image. Then, the difference image shown in Figure 9(C) is obtained based on the maximum value image and the minimum value image. Furthermore, in step S4, N pixel values corresponding to N pixels can be obtained from the difference image shown in Figure 9(C). These pixel values have a one-to-one correspondence with the pixels. The maximum value among the N pixel values is determined as the pixel value to be processed. A segmentation threshold is calculated based on the pixel value to be processed and a proportional threshold. When the pixel value corresponding to a pixel in the difference image is greater than or equal to the segmentation threshold, the pixel is determined as a foreground pixel in the binarized image. When the pixel value corresponding to a pixel in the difference image is less than the segmentation threshold, the pixel is determined as a background pixel in the binarized image, thus obtaining the binarized image shown in Figure 9(D). In step S5, a flooding algorithm is used to detect background regions containing multiple background pixels in the binarized image. Then, based on the binarized image and the background regions in the binarized image, background pixels within the foreground region of the binarized image are obtained. These background pixels within the foreground region of the binarized image are changed to foreground pixels, thus obtaining the hole-filling image shown in Figure 9(E). In step S6, the hole-filling image is subjected to median filtering to obtain a filtered image including the foreground region to be processed. The boundary line of the foreground region to be processed, which includes M pixels, is obtained. For each of the M pixels on the boundary line, K pixels are extended outward to obtain the foreground segmentation result shown in Figure 9(F), where N is an integer greater than 1.

Furthermore, foreground segmentation results can be generated for different medical images to be processed. Please refer to Figure 10, which is a schematic diagram of an embodiment of the foreground segmentation result in this application. As shown in Figure 10, (A) shows a medical image to be processed containing pure white and gray areas. Through the medical image processing method provided in this application, the foreground segmentation result shown in Figure 10 (B) can be obtained. Figure 10 (C) shows a medical image to be processed containing regular vertical stripes. These regular vertical stripes are generated by the scanner scanning the slide. The generation of these regular vertical stripes depends on the scanning device. Then, through the medical image processing method provided in this application, the foreground segmentation result shown in Figure 10 (D) can be obtained. Secondly, Figure 10 (E) shows a medical image to be processed containing black and white stripes. These black and white stripes can be generated by format conversion or by unclear areas generated by the scanner scanning the slide, where black and white stripes are added. Then, through the medical image processing method provided in this application, the foreground segmentation result shown in Figure 10 (F) can be obtained. As can be seen, before binarizing the image, a difference image is generated using the color information of different channels. Since the color information of grayscale pixels differs little in different channels, while the color information of colored pixels differs much in different channels, the color information in the various medical images to be processed shown in Figure 10 can be effectively utilized. The pathological tissue regions extracted based on the difference image are more accurate and have a positive impact on subsequent image analysis.

Based on the above description, the image processing method in this application will be described below. Please refer to Figure 11, which is a schematic diagram of an embodiment of the image processing method in this application. As shown in the figure, an embodiment of the image processing method in this application includes:

201. Obtain a first image to be processed and a second image to be processed, wherein the first image to be processed is a color image, and the first image to be processed includes first image data, second image data and third image data, and the first image data, second image data and third image data respectively correspond to color information under different channels;

In this embodiment, the image processing device can acquire a first image to be processed and a second image to be processed. The first image to be processed may include first image data, second image data, and third image data, and the first image data, second image data, and third image data respectively correspond to color information under different channels. The first image to be processed and the second image to be processed can be images received by the image processing device through a wired network, or images stored by the image processing device itself. Specifically, the first image to be processed is similar to the medical image to be processed described in step 101 above, and will not be repeated here.

It should be understood that in practical applications, the specific color information corresponding to the first image data, the second image data, and the third image data should be flexibly determined based on the actual situation. Furthermore, the image processing device can be deployed on a server or on a terminal device with high computing power; this embodiment uses the deployment of the image processing device on a server as an example.

Specifically, suppose the first image to be processed is a photograph taken on a cloudy day, with a cloudy background and a red car. The second image to be processed is a photograph of a blue sky and a sea.

202. Generate a difference image based on the first image data, the second image data, and the third image data included in the first image to be processed;

In this embodiment, the image processing device can generate a difference image based on the first image data, second image data, and third image data included in the first image to be processed obtained in step 201. Specifically, the difference image is a grayscale image. The method for generating the difference image in this embodiment is similar to that in the embodiment corresponding to FIG2 above, and will not be described again here.

Specifically, the resulting difference image reveals the outline of the car.

203. Perform binarization on the difference image to obtain a binarized image;

In this embodiment, the image processing device can binarize the difference image generated in step 202 to obtain a binarized image. Specifically, this embodiment uses adaptive binarization for foreground segmentation, that is, binarizing the difference image to obtain a binarized image. The method for generating the binarized image in this embodiment is similar to the embodiment corresponding to Figure 2 above, and will not be described again here.

Specifically, the generated binarized image can accurately show the outline of the car.

204. Generate the foreground segmentation result corresponding to the first image to be processed based on the binarized image;

In this embodiment, the image processing device can generate a foreground segmentation result corresponding to the first image to be processed based on the binarized image obtained in step 203. The method for generating the foreground segmentation result in this embodiment is similar to the embodiment corresponding to Figure 2 above, and will not be described again here.

205. Based on the foreground segmentation results, extract the target object from the first image to be processed;

In this embodiment, the image processing device can extract the target object from the first image to be processed based on the foreground segmentation result generated in step 204. If the first image to be processed is a medical image, the target object can be a pathological tissue area. If the first image to be processed is a high-resolution remote sensing image, the target object can be a vegetation area. If the first image to be processed is a real-time traffic monitoring image, the target object can be a bicycle or a car.

Specifically, at this point, the image of the car can be extracted from the first image to be processed, meaning the image of the car is the target object.

206. Generate a composite image based on the target object and the second image to be processed, wherein the target object is located in the first layer, the second image to be processed is located in the second layer, and the first layer covers the second layer.

In this embodiment, the image processing device sets the target object as the first layer, sets the second image to be processed as the second layer, and overlays the first layer on top of the second layer to generate a composite image.

Specifically, an image of a car is overlaid on a photo of blue sky and white clouds to create a composite image. In this composite image, the background of the car is no longer cloudy, but rather blue sky and white clouds.

In this embodiment of the application, an image processing method is provided. Using the above method, since the color information differences between grayscale pixels in different channels are small, while the color information differences between color pixels in different channels are large, a difference image is generated using the color information of different channels before binarizing the image. This effectively utilizes the color information in the image, and the target object extracted based on the difference image is more accurate. Since the layer containing the target object is overlaid on the layer containing the second image to be processed, the resulting composite image accurately summarizes the target object, thereby improving the accuracy of the composite image and positively impacting subsequent image analysis.

The medical image processing apparatus of this application is described in detail below. Please refer to Figure 12, which is a schematic diagram of one embodiment of the medical image processing apparatus in this application. The medical image processing apparatus 300 includes:

The acquisition module 301 is used to acquire a medical image to be processed, wherein the medical image to be processed is a color image, and the medical image to be processed includes first image data, second image data and third image data, and the first image data, second image data and third image data respectively correspond to color information under different channels;

The generation module 302 is used to generate a difference image based on the first image data, the second image data, and the third image data included in the medical image to be processed;

Processing module 303 is used to perform binarization processing on the difference image to obtain a binarized image;

The generation module 302 is also used to generate the foreground segmentation result corresponding to the medical image to be processed based on the binarized image.

In this embodiment of the application, a method for medical image processing is provided. In this method, since the color information difference of grayscale pixels in different channels is small, while the color information difference of color pixels in different channels is large, before binarizing the image, a difference image is generated using the color information of different channels. This effectively utilizes the color information in the image, and the pathological tissue region extracted based on the difference image is more accurate, which has a positive impact on subsequent image analysis.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing apparatus 300 provided in this application,

The generation module 302 is specifically used to generate a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed.

A minimum value image is generated based on the first image data, the second image data, and the third image data included in the medical image to be processed.

Generate a difference image based on the maximum and minimum value images.

In this embodiment of the application, a method for generating a difference image is provided. By means of the above method, a maximum value image and a minimum value image are generated based on the first image data, the second image data, and the third image data. Since the color information corresponding to different image data is different, the maximum value image and the minimum value image determined based on different image data include color information of the medical image to be processed with high accuracy, thereby improving the accuracy of the difference image generation.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing apparatus 300 provided in this application,

The generation module 302 is specifically used to determine the maximum pixel value corresponding to the target pixel from the first pixel value, the second pixel value, and the third pixel value according to the first image data, the second image data, and the third image data included in the medical image to be processed. The first pixel value is the pixel value corresponding to the first pixel position in the first image data, the second pixel value is the pixel value corresponding to the second pixel position in the second image data, the third pixel value is the pixel value corresponding to the third pixel position in the third image data, and the target pixel is the pixel value corresponding to the fourth pixel position in the maximum value image. The first pixel position, the second pixel position, the third pixel position, and the fourth pixel position all correspond to the position of the same pixel in the medical image to be processed.

The generation module 302 is specifically used to determine the minimum pixel value corresponding to the target pixel from the first pixel value, the second pixel value, and the third pixel value based on the first image data, the second image data, and the third image data included in the medical image to be processed.

The generation module 302 is specifically used to subtract the maximum pixel value corresponding to the target pixel in the maximum value image from the minimum pixel value corresponding to the target pixel in the maximum value image to obtain the difference pixel value corresponding to the target pixel in the difference image.

In this embodiment of the application, a method for generating a maximum value image is provided. In the above manner, the maximum pixel value and the minimum pixel value are determined by the pixel values of the target pixel points corresponding to the first image data, the second image data, and the third image data. The maximum pixel value and the minimum pixel value reflect the color information of the medical image to be processed to different degrees. The difference pixel value is obtained by subtracting the maximum pixel value and the minimum pixel value, so that the difference pixel value can accurately reflect the color information of the medical image to be processed, thereby improving the accuracy of the difference image generation.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing apparatus 300 provided in this application,

The generation module 302 is specifically used to generate a difference image to be processed based on the first image data, the second image data, and the third image data included in the medical image to be processed.

Gaussian blur is applied to the difference image to obtain the difference image.

In this embodiment of the application, another method for generating a difference image is provided. In this method, the difference image to be processed is subjected to Gaussian blur processing. Since Gaussian blur processing can improve segmentation robustness, the resulting difference image has better segmentation robustness, thereby improving the stability of the difference image.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing device 300 provided in this application, the medical image processing device 300 further includes a determination module 304;

The determination module 304 is used to determine the segmentation threshold based on the difference image;

The determining module 304 is further configured to determine the pixel as the foreground pixel of the binarized image if the pixel value corresponding to the pixel in the difference image is greater than or equal to the segmentation threshold.

The determination module 304 is also used to determine the pixel as the background pixel of the binarized image if the pixel value corresponding to the pixel in the difference image is less than the segmentation threshold.

In this embodiment of the application, a method for obtaining a binarized image is provided. By means of the above method, a binarized image is generated according to the binarization process. Since the geometric properties of the binarized image do not involve the gray values of the pixels, the subsequent processing of the binarized image can be simplified, thereby improving the efficiency of generating foreground segmentation results.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing apparatus 300 provided in this application,

The determination module 304 is specifically used to obtain N pixel values corresponding to N pixels based on the difference image, wherein there is a one-to-one correspondence between pixel values and pixel points, and N is an integer greater than 1;

Determine the pixel value to be processed from N pixel values, where the pixel value to be processed is the maximum value among the N pixel values;

The segmentation threshold is calculated based on the pixel value to be processed and the ratio threshold.

In this embodiment of the application, another method for obtaining the segmentation threshold is provided. In the above manner, the segmentation threshold can be determined by the pixel value to be processed by the maximum pixel value and the ratio threshold. Since the gray depth of the difference image is different and the brightness distribution of different regions can also be different, the segmentation threshold can be flexibly determined by adjusting the ratio threshold, thereby improving the accuracy and flexibility of the threshold and thus improving the accuracy of binarized image generation.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing apparatus 300 provided in this application,

The generation module 302 is specifically used to detect the background region in the binarized image using a flooding algorithm, wherein the background region includes multiple background pixels;

Based on the binarized image and the background region in the binarized image, obtain the background pixels within the foreground region of the binarized image, wherein the foreground region includes multiple foreground pixels;

By replacing background pixels in the foreground region of a binarized image with foreground pixels, a hole-filling image is obtained.

Median filtering is applied to the hole-filling image to obtain the foreground segmentation result corresponding to the medical image to be processed.

In this embodiment of the application, a method for generating foreground segmentation results is provided. By changing the background pixels in the foreground region to foreground pixels in the above manner, the resulting hole-filling image has good reliability. Secondly, by median filtering, the foreground segmentation result corresponding to the medical image to be processed can be clear and have good visual effect without damaging the contour and edge features of the image.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing apparatus 300 provided in this application,

The processing module 303 is specifically used to perform median filtering on the hole-filling image to obtain a filtered image, wherein the filtered image includes the foreground region to be processed;

Obtain the boundary line of the foreground region to be processed, where the boundary line consists of M pixels, M being an integer greater than 1;

For each of the M pixels on the boundary line, extend outward by K pixels to obtain the foreground segmentation result, where K is an integer greater than or equal to 1.

This application provides another method for generating foreground segmentation results. Through median filtering, the filtered image is clear and has good visual quality without damaging the image's contours and edges. Furthermore, a flooding algorithm is used to perform morphological processing on the filtered image, improving the accuracy and uniformity of the foreground segmentation results.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing apparatus 300 provided in this application,

Acquisition module 301 is specifically used to acquire raw medical images;

A sliding window method is used to extract medical sub-images from the original medical image;

If a pathological tissue region is detected in a medical sub-image, it is identified as a medical image to be processed.

If a medical sub-image is found to not include a pathological tissue region, the medical sub-image is identified as a background image and removed.

This application provides a method for acquiring a medical image to be processed. By detecting whether a medical sub-image includes a pathological tissue region, the medical image to be processed is determined. This ensures that the medical image containing a pathological tissue region can be segmented into a foreground segmentation result through the aforementioned steps. Furthermore, the foreground segmentation result includes the pathological tissue region, facilitating subsequent processing and analysis of this pathological tissue region. Secondly, the medical sub-image containing the pathological tissue region is designated as a background image and removed to reduce resource consumption.

Optionally, based on the embodiment corresponding to FIG12 above, in another embodiment of the medical image processing device 300 provided in this application, the image processing device further includes a training module 305;

The generation module 302 is also used to generate a target positive sample image based on the foreground segmentation result, wherein the target positive sample image belongs to a positive sample image in the positive sample set, and each positive sample image contains a pathological tissue region;

The acquisition module 301 is also used to acquire a negative sample set, wherein the negative sample set includes at least one negative sample image, and each negative sample image does not contain a pathological tissue region;

Training module 305 is used to train the image segmentation model based on the positive sample set and the negative sample set.

In this embodiment of the application, a method for training an image segmentation model is provided. By using the above method, the image segmentation model is trained with a set of positive sample images containing pathological tissue regions and a set of negative sample images not containing pathological tissue regions, thereby improving the accuracy and reliability of the image segmentation model and thus improving the efficiency and accuracy of image processing.

The image processing apparatus of this application is described in detail below. Please refer to FIG13, which is a schematic diagram of an embodiment of the image processing apparatus in this application. The image processing apparatus 400 includes:

The acquisition module 401 is used to acquire a first image to be processed and a second image to be processed, wherein the first image to be processed is a color image and includes first image data, second image data and third image data, and the first image data, second image data and third image data respectively correspond to color information under different channels;

The generation module 402 is used to generate a difference image based on the first image data, the second image data, and the third image data included in the first image to be processed;

Processing module 403 is used to perform binarization processing on the difference image to obtain a binarized image;

The generation module 402 is also used to generate a foreground segmentation result corresponding to the first image to be processed based on the binarized image;

Extraction module 404 is used to extract the target object from the first image to be processed based on the foreground segmentation result;

The generation module 402 is also used to generate a composite image based on the target object and the second image to be processed, wherein the target object is located in the first layer, the second image to be processed is located in the second layer, and the first layer covers the second layer.

In this embodiment of the application, an image processing method is provided. Using the above method, since the color information differences between grayscale pixels in different channels are small, while the color information differences between color pixels in different channels are large, a difference image is generated using the color information of different channels before binarizing the image. This effectively utilizes the color information in the image, and the target object extracted based on the difference image is more accurate. Since the layer containing the target object is overlaid on the layer containing the second image to be processed, the resulting composite image accurately summarizes the target object, thereby improving the accuracy of the composite image and positively impacting subsequent image analysis.

Figure 14 is a schematic diagram of a server structure provided in an embodiment of this application. The server 500 can vary significantly due to different configurations or performance. It may include one or more central processing units (CPUs) 522 (e.g., one or more processors) and memory 532, and one or more storage media 530 (e.g., one or more mass storage devices) for storing application programs 542 or data 544. The memory 532 and storage media 530 can be temporary or persistent storage. The program stored in the storage media 530 may include one or more modules (not shown in the figure), each module including a series of instruction operations on the server. Furthermore, the CPU 522 may be configured to communicate with the storage media 530 and execute the series of instruction operations in the storage media 530 on the server 500.

Server 500 may also include one or more power supplies 525, one or more wired or wireless network interfaces 550, one or more input/output interfaces 558, and/or one or more operating systems 541, such as Windows Server ^™ , Mac OS X ^™ , Unix ^™ , Linux ^™ , FreeBSD ^™ , etc.

The steps performed by the server in the above embodiments can be based on the server structure shown in Figure 14.

In this embodiment, CPU 522 is used to execute the steps performed by the medical image processing device in the embodiment corresponding to FIG2, and CPU 522 is also used to execute the steps performed by the image processing device in the embodiment corresponding to FIG1.

Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims (11)

1. A method for medical image processing, characterized in that it includes: Acquire a medical image to be processed, wherein the medical image to be processed is a color image, and the medical image to be processed includes first image data, second image data and third image data, and the first image data, the second image data and the third image data respectively correspond to color information in different channels in the same medical image to be processed; A difference image is generated based on the first image data, the second image data, and the third image data included in the medical image to be processed; The difference image is binarized to obtain a binarized image; Generate the foreground segmentation result corresponding to the medical image to be processed based on the binarized image; The step of generating a difference image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: generating a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed; generating a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed; and generating the difference image based on the maximum value image and the minimum value image. The step of generating a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: determining the maximum pixel value corresponding to a target pixel from the first pixel value, the second pixel value, and the third pixel value included in the first image data, the second image data, and the third image data, wherein the first pixel value is the pixel value corresponding to the first pixel position in the first image data, the second pixel value is the pixel value corresponding to the second pixel position in the second image data, the third pixel value is the pixel value corresponding to the third pixel position in the third image data, and the target pixel is the pixel value corresponding to the fourth pixel position in the maximum value image, wherein the first pixel position, the second pixel position, and the third pixel position are... The fourth pixel position corresponds to the position of the same pixel in the medical image to be processed; generating a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: determining the minimum pixel value corresponding to the target pixel from the first pixel value, the second pixel value, and the third pixel value based on the first image data, the second image data, and the third image data included in the medical image to be processed; generating a difference image based on the maximum value image and the minimum value image includes: subtracting the maximum pixel value corresponding to the target pixel in the maximum value image from the minimum pixel value corresponding to the target pixel in the minimum value image to obtain the difference pixel value corresponding to the target pixel in the difference image; The step of binarizing the difference image to obtain a binarized image includes: determining a segmentation threshold using dynamic threshold binarization based on the difference image; if the pixel value corresponding to a pixel in the difference image is greater than or equal to the segmentation threshold, then the pixel is determined as a foreground pixel of the binarized image; if the pixel value corresponding to a pixel in the difference image is less than the segmentation threshold, then the pixel is determined as a background pixel of the binarized image. The step of determining the segmentation threshold using dynamic threshold binarization based on the difference image includes: obtaining N pixel values corresponding to N pixels based on the difference image, wherein the pixel values and the pixels have a one-to-one correspondence, and N is an integer greater than 1; determining the pixel value to be processed from the N pixel values, wherein the pixel value to be processed is the maximum value among the N pixel values; and multiplying the pixel value to be processed by a proportional threshold to calculate the segmentation threshold. 2. The method according to claim 1, characterized in that generating a difference image based on the first image data, the second image data, and the third image data included in the medical image to be processed comprises: A difference image to be processed is generated based on the first image data, the second image data, and the third image data included in the medical image to be processed; The difference image to be processed is subjected to Gaussian blurring to obtain the difference image. 3. The method according to claim 1, characterized in that, generating the foreground segmentation result corresponding to the medical image to be processed based on the binarized image includes: A flooding algorithm is used to detect the background region in the binarized image, wherein the background region includes multiple background pixels; Based on the binarized image and the background region in the binarized image, background pixels within the foreground region of the binarized image are obtained, wherein the foreground region includes multiple foreground pixels; By changing the background pixels in the foreground region of the binarized image to foreground pixels, a hole-filling image is obtained. The hole-filling image is subjected to median filtering to obtain the foreground segmentation result corresponding to the medical image to be processed. 4. The method according to claim 3, characterized in that, the step of performing median filtering on the hole-filling image to obtain the foreground segmentation result corresponding to the medical image to be processed includes: The hole-filling image is subjected to median filtering to obtain a filtered image, wherein the filtered image includes the foreground region to be processed; Obtain the boundary line of the foreground region to be processed, wherein the boundary line includes M pixels, and M is an integer greater than 1; For each of the M pixels on the boundary line, extend outward by K pixels to obtain the foreground segmentation result, where K is an integer greater than or equal to 1. 5. The method according to claim 1, wherein acquiring the medical image to be processed includes: Acquire raw medical images; A sliding window is used to extract medical sub-images from the original medical image; If a pathological tissue region is detected in the medical sub-image, it is identified as the medical image to be processed; If the medical sub-image does not include a pathological tissue area, then the medical sub-image is identified as a background image and the background image is removed. 6. The method according to any one of claims 1 to 5, characterized in that, after generating the foreground segmentation result corresponding to the medical image to be processed based on the binarized image, the method further includes: A target positive sample image is generated based on the foreground segmentation result, wherein the target positive sample image belongs to a positive sample image in the positive sample set, and each positive sample image contains a pathological tissue region; Obtain a negative sample set, wherein the negative sample set includes at least one negative sample image, and each negative sample image does not contain a pathological tissue region; The image segmentation model is trained based on the set of positive samples and the set of negative samples. 7. An image processing method, characterized in that it comprises: Acquire a first image to be processed and a second image to be processed, wherein the first image to be processed is a color image, and the first image to be processed includes first image data, second image data and third image data, and the first image data, the second image data and the third image data respectively correspond to the color information of different channels in the same medical image to be processed; A difference image is generated based on the first image data, the second image data, and the third image data included in the first image to be processed; The difference image is binarized to obtain a binarized image; Generate the foreground segmentation result corresponding to the first image to be processed based on the binarized image; Based on the foreground segmentation results, the target object is extracted from the first image to be processed; A composite image is generated based on the target object and the second image to be processed, wherein the target object is located in a first layer, the second image to be processed is located in a second layer, and the first layer covers the second layer; The step of generating a difference image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: generating a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed; generating a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed; and generating the difference image based on the maximum value image and the minimum value image. The step of generating a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: determining the maximum pixel value corresponding to a target pixel from the first pixel value, the second pixel value, and the third pixel value included in the first image data, the second image data, and the third image data, wherein the first pixel value is the pixel value corresponding to the first pixel position in the first image data, the second pixel value is the pixel value corresponding to the second pixel position in the second image data, the third pixel value is the pixel value corresponding to the third pixel position in the third image data, and the target pixel is the pixel value corresponding to the fourth pixel position in the maximum value image, wherein the first pixel position, the second pixel position, and the third pixel position are... The fourth pixel position corresponds to the position of the same pixel in the medical image to be processed; generating a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: determining the minimum pixel value corresponding to the target pixel from the first pixel value, the second pixel value, and the third pixel value based on the first image data, the second image data, and the third image data included in the medical image to be processed; generating a difference image based on the maximum value image and the minimum value image includes: subtracting the maximum pixel value corresponding to the target pixel in the maximum value image from the minimum pixel value corresponding to the target pixel in the minimum value image to obtain the difference pixel value corresponding to the target pixel in the difference image; The step of binarizing the difference image to obtain a binarized image includes: determining a segmentation threshold using dynamic threshold binarization based on the difference image; if the pixel value corresponding to a pixel in the difference image is greater than or equal to the segmentation threshold, then the pixel is determined as a foreground pixel of the binarized image; if the pixel value corresponding to a pixel in the difference image is less than the segmentation threshold, then the pixel is determined as a background pixel of the binarized image. The step of determining the segmentation threshold using dynamic threshold binarization based on the difference image includes: obtaining N pixel values corresponding to N pixels based on the difference image, wherein the pixel values and the pixels have a one-to-one correspondence, and N is an integer greater than 1; determining the pixel value to be processed from the N pixel values, wherein the pixel value to be processed is the maximum value among the N pixel values; and multiplying the pixel value to be processed by a proportional threshold to calculate the segmentation threshold. 8. A medical image processing device, characterized in that it comprises: The acquisition module is used to acquire a medical image to be processed, wherein the medical image to be processed is a color image, and the medical image to be processed includes first image data, second image data and third image data, and the first image data, the second image data and the third image data respectively correspond to color information in different channels in the same medical image to be processed; The generation module is used to generate a difference image based on the first image data, the second image data, and the third image data included in the medical image to be processed; The processing module is used to perform binarization processing on the difference image to obtain a binarized image; The generation module is also used to generate a foreground segmentation result corresponding to the medical image to be processed based on the binarized image; The generation module is further configured to generate a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed; generate a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed; and generate the difference image based on the maximum value image and the minimum value image. The step of generating a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: determining the maximum pixel value corresponding to a target pixel from the first pixel value, the second pixel value, and the third pixel value included in the first image data, the second image data, and the third image data, wherein the first pixel value is the pixel value corresponding to the first pixel position in the first image data, the second pixel value is the pixel value corresponding to the second pixel position in the second image data, the third pixel value is the pixel value corresponding to the third pixel position in the third image data, and the target pixel is the pixel value corresponding to the fourth pixel position in the maximum value image, wherein the first pixel position, the second pixel position, and the third pixel position are... The fourth pixel position corresponds to the position of the same pixel in the medical image to be processed; generating a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: determining the minimum pixel value corresponding to the target pixel from the first pixel value, the second pixel value, and the third pixel value based on the first image data, the second image data, and the third image data included in the medical image to be processed; generating a difference image based on the maximum value image and the minimum value image includes: subtracting the maximum pixel value corresponding to the target pixel in the maximum value image from the minimum pixel value corresponding to the target pixel in the minimum value image to obtain the difference pixel value corresponding to the target pixel in the difference image; The processing module is specifically used to determine a segmentation threshold based on the difference image using dynamic threshold binarization; if the pixel value corresponding to a pixel in the difference image is greater than or equal to the segmentation threshold, then the pixel is determined as a foreground pixel in the binarized image; if the pixel value corresponding to a pixel in the difference image is less than the segmentation threshold, then the pixel is determined as a background pixel in the binarized image. The step of determining the segmentation threshold using dynamic threshold binarization based on the difference image includes: obtaining N pixel values corresponding to N pixels based on the difference image, wherein the pixel values and the pixels have a one-to-one correspondence, and N is an integer greater than 1; determining the pixel value to be processed from the N pixel values, wherein the pixel value to be processed is the maximum value among the N pixel values; and multiplying the pixel value to be processed by a proportional threshold to calculate the segmentation threshold. 9. An image processing apparatus, characterized in that it comprises: The acquisition module is used to acquire a first image to be processed and a second image to be processed, wherein the first image to be processed is a color image, and the first image to be processed includes first image data, second image data and third image data, and the first image data, the second image data and the third image data respectively correspond to the color information of different channels in the same medical image to be processed; The generation module is used to generate a difference image based on the first image data, the second image data, and the third image data included in the first image to be processed; The processing module is used to perform binarization processing on the difference image to obtain a binarized image; The generation module is further configured to generate a foreground segmentation result corresponding to the first image to be processed based on the binarized image; The extraction module is used to extract the target object from the first image to be processed based on the foreground segmentation result; The generation module is further configured to generate a composite image based on the target object and the second image to be processed, wherein the target object is located in a first layer, the second image to be processed is located in a second layer, and the first layer covers the second layer; The generation module is further configured to generate a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed; generate a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed; and generate the difference image based on the maximum value image and the minimum value image. The step of generating a maximum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: determining the maximum pixel value corresponding to a target pixel from the first pixel value, the second pixel value, and the third pixel value included in the first image data, the second image data, and the third image data, wherein the first pixel value is the pixel value corresponding to the first pixel position in the first image data, the second pixel value is the pixel value corresponding to the second pixel position in the second image data, the third pixel value is the pixel value corresponding to the third pixel position in the third image data, and the target pixel is the pixel value corresponding to the fourth pixel position in the maximum value image, wherein the first pixel position, the second pixel position, and the third pixel position are... The fourth pixel position corresponds to the position of the same pixel in the medical image to be processed; generating a minimum value image based on the first image data, the second image data, and the third image data included in the medical image to be processed includes: determining the minimum pixel value corresponding to the target pixel from the first pixel value, the second pixel value, and the third pixel value based on the first image data, the second image data, and the third image data included in the medical image to be processed; generating a difference image based on the maximum value image and the minimum value image includes: subtracting the maximum pixel value corresponding to the target pixel in the maximum value image from the minimum pixel value corresponding to the target pixel in the minimum value image to obtain the difference pixel value corresponding to the target pixel in the difference image; The processing module is specifically used to determine a segmentation threshold based on the difference image using dynamic threshold binarization; if the pixel value corresponding to a pixel in the difference image is greater than or equal to the segmentation threshold, then the pixel is determined as a foreground pixel in the binarized image; if the pixel value corresponding to a pixel in the difference image is less than the segmentation threshold, then the pixel is determined as a background pixel in the binarized image. The step of determining the segmentation threshold using dynamic threshold binarization based on the difference image includes: obtaining N pixel values corresponding to N pixels based on the difference image, wherein the pixel values and the pixels have a one-to-one correspondence, and N is an integer greater than 1; determining the pixel value to be processed from the N pixel values, wherein the pixel value to be processed is the maximum value among the N pixel values; and multiplying the pixel value to be processed by a proportional threshold to calculate the segmentation threshold. 10. A computer device, characterized in that it comprises: a memory, a transceiver, a processor, and a bus system; The memory is used to store programs; The processor is configured to execute a program in the memory to implement the method as described in any one of claims 1 to 6 above, or to implement the method as described in claim 7 above; The bus system is used to connect the memory and the processor to enable communication between the memory and the processor. 11. A computer-readable storage medium comprising instructions, when executed on a computer, causing the computer to perform the method as claimed in any one of claims 1 to 6, or to perform the method as claimed in claim 7.