CN111986165B - Calcification detection method and device in breast image - Google Patents
Calcification detection method and device in breast image Download PDFInfo
- Publication number
- CN111986165B CN111986165B CN202010758673.9A CN202010758673A CN111986165B CN 111986165 B CN111986165 B CN 111986165B CN 202010758673 A CN202010758673 A CN 202010758673A CN 111986165 B CN111986165 B CN 111986165B
- Authority
- CN
- China
- Prior art keywords
- calcification
- image
- calcified
- detection
- type
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30068—Mammography; Breast
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Radiology & Medical Imaging (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a calcification detection method and device in breast images, computer equipment and a computer readable storage medium. The method comprises the following steps: the breast image is preprocessed to obtain a first image. And inputting the first image into N calcification detection models to obtain calcification regions and calcification types thereof in the first image respectively, wherein N is a positive integer. And fusing the calcified region and the calcification type thereof in the obtained first image based on a preset rule to obtain a calcification detection result. According to the scheme, the efficiency and the accuracy of calcification detection in breast images are improved.
Description
Technical Field
The present invention relates to the field of medical technology, and in particular, to a method and apparatus for detecting calcification in breast images, a computer device, and a computer readable storage medium.
Background
The breast molybdenum target examines the female breast by using low dose X-ray, thereby diagnosing various diseases related to female breast, such as breast tumor, cyst, etc. At present, the breast molybdenum target is mainly applied to screening of early breast cancer, so that effective detection of early manifestations of various breast cancers on a breast molybdenum target image is important. In general, diagnosis of breast lesions is often based on detection of lesion symptoms, including: calcification (malignant large calcification, malignant small calcification, benign calcification), nipple retraction, skin retraction, tumor/asymmetry, structural distortion, etc.
In the prior art, doctors usually diagnose the breast molybdenum target image through personal experience, the efficiency is low, and the subjectivity exists.
Therefore, how to determine the lesion symptoms in the breast image and improve the efficiency and accuracy of identifying the lesion symptoms becomes one of the problems to be solved in the present day.
Disclosure of Invention
The invention provides a calcification detection method, a calcification detection device, a calcification detection computer device and a calcification detection computer readable storage medium for detecting calcifications of different calcification types in a breast image. The efficiency and accuracy of calcification detection in breast images are improved, and the diagnosis accuracy of the subsequent diagnosis based on the calcification detection is further improved.
The invention provides a calcification detection method in breast images, which comprises the following steps:
preprocessing the breast image to obtain a first image;
inputting the first image into N calcification detection models to respectively obtain calcification regions and calcification types of the calcification regions in the first image, wherein N is a positive integer;
and fusing the calcified region and the calcification type thereof in the obtained first image based on a preset rule to obtain a calcification detection result.
Optionally, the preprocessing the breast image to obtain a first image includes: and binarizing the breast image, and taking the area with the largest area in the binarized breast image as a first image.
Optionally, based on a preset rule, fusing the calcified region and the calcification type thereof in the obtained first image to obtain a calcification detection result includes: the calcification detection result is obtained based on the overlap condition of the first calcified region and the second calcified region and the calcification type.
Optionally, the obtaining the calcification detection result based on the overlapping condition of the first calcified region and the second calcified region and the calcification type includes:
and determining the intersection ratio of the first calcified region and the second calcified region, and if the intersection ratio is larger than a preset value, merging the first calcified region and the calcified type thereof and the second calcified region and the calcified type thereof based on a merging rule to obtain a calcification detection result.
Optionally, determining the intersection ratio of the first calcified region and the second calcified region comprises:
determining the number of first pixel points covered by the intersection of the first calcified region and the second calcified region;
determining the number of second pixel points covered by a union of the first calcified region and the second calcified region;
and determining the ratio of the number of the first pixel points to the number of the second pixel points as the intersection ratio of the first calcified region and the second calcified region.
Optionally, the calcification types include: a malignant large calcification type and a malignant small calcification type, the merge rule comprising: combining the malignant large calcification type and the malignant small calcification type into a malignant large calcification type.
Optionally, the calcification detection model is obtained by respectively training pre-classified training samples of each calcification type, and calcification areas in the training samples are obtained by the following modes:
binarizing the marked calcified region to obtain a circumscribed rectangular frame of the binarized marked calcified region;
and taking the area where the circumscribed rectangular frame is positioned as a calcified area.
Optionally, the calcification types include: malignant large calcification type, malignant small calcification type, and benign calcification type.
The present invention also provides a calcification detection apparatus in breast image, comprising:
a preprocessing unit for preprocessing the breast image to obtain a first image;
n calcification detection models are used for inputting the first image and respectively outputting calcification areas and calcification types of the calcification areas in the first image, wherein N is a positive integer;
and the fusion unit is used for fusing the calcified region and the calcification type thereof in the obtained first image based on a preset rule to obtain a calcification detection result.
The invention also provides a computer device comprising at least one processor and at least one memory, wherein the memory stores a computer program which, when executed by the processor, enables the processor to perform the above-described method of calcification detection in breast images.
The invention also provides a computer readable storage medium, which when executed by a processor within the device, causes the device to perform the above-described method of calcification detection in breast images.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
the breast image is preprocessed to obtain a first image. And inputting the first image into N calcification detection models to obtain calcification regions and calcification types thereof in the first image respectively, wherein N is a positive integer. And fusing the calcified region and the calcification type thereof in the obtained first image based on a preset rule to obtain a calcification detection result. Since the calcification foci present in the breast image are no longer determined manually, the efficiency and accuracy of calcification detection in the breast image is improved. In addition, the calcification detection is carried out in a targeted manner through the calcification detection models of different calcification types, so that the efficiency and accuracy of calcification detection are improved to a certain extent. In addition, the accuracy of calcification detection is further improved by fusing the detected calcified region and the calcified type thereof.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
FIG. 1 is a flow chart of a method for detecting calcification in breast images according to an embodiment of the invention;
fig. 2 is a schematic structural view of a calcification detecting apparatus in breast image according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As mentioned in the background art, the prior art is time-consuming, laborious and highly subjective, in which the presence and type of calcification in the breast image is determined manually. For calcification, it is generally classified into malignant calcification and benign calcification, and malignant calcification is classified into malignant large calcification and malignant small calcification. The inventor considers that when detecting calcification, if a single calcification detection model is adopted, missed detection and detection efficiency may be low, so that it is proposed to detect calcification with pertinence through different types of calcification detection models, in addition, in order to further improve the accuracy of calcification detection, after detecting calcification areas through different calcification types of calcification detection models, judgment is performed according to preset rules so as to combine the calcification areas detected by different calcification types of calcification detection models and calcification types to obtain a final calcification detection result. In addition, in order to improve the performance of the calcification detection model, when the calcification detection model is trained, the calcification region marking in the training sample is performed, the calcification region marking is performed on the basis of the manually marked calcification region, and finally, the binarized circumscribed rectangular frame of the manually marked calcification region is used as the calcification region.
Fig. 1 is a flow chart of a method for detecting calcification in breast images according to an embodiment of the invention. As shown in fig. 1, the calcification detection method in the breast image includes:
s11: the breast image is preprocessed to obtain a first image.
S12: and inputting the first image into N calcification detection models to obtain calcification regions and calcification types thereof in the first image respectively, wherein N is a positive integer.
S13: and fusing the calcified region and the calcification type thereof in the obtained first image based on a preset rule to obtain a calcification detection result.
S11 is performed, preprocessing the breast image to obtain a first image. In this embodiment, the preprocessing includes: and binarizing the breast image, and taking the area with the largest area in the binarized breast image as a first image. Specifically, the binarized threshold value may be obtained by a maximum class pitch method of a gray level histogram of a preset region in the breast image, and the gray level value of a pixel point in the breast image is converted into 0 or 1 by the binarization operation, respectively. The first image obtained by binarization in this step is a breast area in the breast image, and in other embodiments, the first image may be obtained by segmenting the breast area in the breast image by an image segmentation algorithm, such as a watershed algorithm.
S12 is executed: and inputting the first image into N calcification detection models to obtain calcification regions and calcification types thereof in the first image respectively, wherein N is a positive integer.
In this embodiment, the calcification types may include: the calcification model may be a benign calcification detection model, a malignant large calcification detection model, and a malignant small calcification detection model, respectively. I.e. N is equal to 3. By inputting the first image to a different type of calcification detection model, if it is input to a malignant large calcification detection model, it is possible to output whether or not the calcification in the first image is a malignant large calcification, and when it is determined that the calcification is a malignant large calcification, the region in which the calcification is located is output. The calcified areas detected by the calcification detection model are marked in the form of rectangular frames, and of course, in other embodiments, the detected calcified areas can be marked in other shapes such as circles, squares, and the like. In this embodiment, after the calcified regions are detected in the first image by a single calcified detection model, the first images marked with the detection frames detected by each calcified detection model may be superimposed to obtain all the calcified regions in the first image.
In this embodiment, the different types of calcification detection models may respectively use a plurality of breast images marked with different types of calcification areas as training samples, or may perform enhancement operations on the plurality of breast images marked with different types of calcification areas, so as to enlarge the data volume of the training samples, where the enhancement operations include: randomly shifting preset pixels (such as 0-30 pixels), randomly rotating a set angle (such as-15 degrees), randomly zooming a set multiple (such as 0.85-1.15 times), and performing small amount of dithering on the contrast and brightness of the breast image.
In order to improve the performance of the calcification detection model in this embodiment, when the calcification detection model is trained, the calcification region marked in the training sample is not directly used as the calcification region, but the calcification region marked by the doctor or the professional labeling personnel is binarized first, then a circumscribed rectangular frame of the binarized marked calcification region is obtained, and the region where the circumscribed rectangular frame is located is used as the calcification region in the training sample. Because the calcified areas marked manually in the training sample are normalized, the accuracy of the calcified detection model can be improved to a great extent when the calcified areas are detected by adopting the calcified detection model.
And inputting a training sample into a corresponding initial calcification detection model, calculating a loss function according to the marked calcification region and the type thereof and the calcification region detected by the initial calcification detection model and the type thereof during training, and repeatedly iterating by adopting a back propagation algorithm and a random gradient descent (SGD, stochastic Gradient Descent) optimization algorithm to continuously update the parameters of the initial calcification detection model. If the loss function of a certain training is smaller than or equal to the threshold value, the initial calcification detection model corresponding to the model parameter of the training can be used as a calcification detection model. After the calcification detection model is obtained by training, the breast image may then be input to the calcification detection model to obtain the calcified region and its type in the breast image.
In this embodiment, specifically, the calcification detection model may include: and the characteristic extraction module and the detection frame acquisition module. The feature extraction module may include: n convolution modules, N max pooling layers, N2D deconvolution layers of 2 gamma layers, and tensor superposition layers. The output of the convolution module is connected with the input of the maximum pooling layer, and the convolution module and the maximum pooling layer are alternately connected. Each convolution module comprises a plurality of convolution blocks, wherein each convolution block comprises: a convolution layer (Conv 2 d), a batch normalization layer (BN, batch Normalization) and an activation layer, the activation function may be a linear rectification function (ReLU, recified Linear Unit). The tensor superimposing layer is used for superimposing tensors of two same dimensions (tensors of one dimension can be output by the pooling layer and tensors of one dimension can be output by the deconvolution layer) at corresponding positions to obtain the feature map. The feature extraction module outputs the feature image output by the tensor superimposed layer and the feature image output by the convolution module to the detection frame acquisition module, so that a plurality of detection frames are detected in the breast image, the detection frames with confidence coefficient larger than a preset threshold value are reserved, then the detection frames with larger overlapping degree are removed through a non-maximum value suppression (Non Maximum Suppression) method, and a final detection result is obtained. For example, if the calcification detection model is a benign calcification detection model, after the breast image is input to the benign calcification detection model, the benign calcification detection model outputs a detection frame belonging to a calcified region in the breast image, and the type of the calcified region is benign calcification or not.
In this embodiment, the feature extraction module may be a feature pyramid network (FPN, momenta Paper Reading), and the detection frame acquisition module may be a SSD (Single Shot MultiBox Detector) network.
To this end, the calcified regions in the breast image and their corresponding calcification types are detected by the different types of calcification detection models described above. In this embodiment, the calcification type includes: the benign calcification type, the malignant large calcification type and the malignant small calcification type are correspondingly described as examples, and in practical application, the calcification types can be divided into different types according to practical requirements to establish calcification detection models corresponding to the different types of calcification types, so the calcification types comprise: benign calcification type, malignant large calcification type and malignant small calcification type should not be taken as limitations of the technical solution of the present invention.
In this embodiment, after the calcified regions and the types thereof are detected by the different types of calcification detection models, in order to further improve the accuracy of calcification detection, the calcified regions and the calcified types detected by the different types of calcification detection models may be determined according to a preset rule, so as to combine to obtain a final calcification detection result. Namely: s13 is performed to fuse the calcified region and its calcification type in the obtained first image to obtain a calcification detection result.
In the present embodiment, the calcification detection result may be obtained based on the overlap condition of the first calcified region and the second calcified region and the calcification type. When determining whether or not the first calcified region and the second calcified region overlap, the determination may be made by the intersection ratio of the first region and the second region. Specifically, the intersection ratio of the first calcified region and the second calcified region is determined, that is: the ratio of the area of intersection of the first and second calcified regions to the area of union of the first and second calcified regions. And if the merging ratio is larger than a preset value, merging the first calcified region and the calcification type thereof and the second calcified region and the calcification type thereof based on a merging rule to obtain a calcification detection result. The preset value may be set according to practical applications, and in this embodiment, the preset value may be 0.5.
In the present embodiment, the intersection ratio of the first calcified region and the second calcified region can be obtained by:
first, the number of first pixels covered by the intersection of the first calcified region and the second calcified region is determined. Then, the number of second pixels covered by the union of the first calcified region and the second calcified region is determined. And determining the ratio of the number of the first pixel points to the number of the second pixel points as the intersection ratio of the first calcified region and the second calcified region.
In this embodiment, after determining that the intersection ratio of the first calcified region and the second calcified region is greater than the preset value, different calcification types may be merged according to the merging rule, specifically, the malignant large calcification type and the malignant small calcification type may be merged into the malignant large calcification type, and the union of the first calcified region and the second calcified region may be used as the calcified region obtained by final detection.
The present invention also provides a calcification detection apparatus in breast images, the apparatus comprising:
a preprocessing unit 10 for preprocessing the breast image to obtain a first image.
And N calcification detection models 11 are used for inputting the first image and respectively outputting calcification areas and calcification types thereof in the first image, wherein N is a positive integer.
And a fusion unit 12, configured to fuse the calcified region and the calcification type thereof in the obtained first image based on a preset rule to obtain a calcification detection result.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a calcification detection apparatus in breast image according to an embodiment of the present invention, and N calcification detection models 11 in fig. 2 include a first calcification detection model, a second calcification detection model, and a third calcification detection model … … nth calcification detection model. Specifically, in this embodiment, N may be 3, the first calcification detection model may be a benign calcification detection model, the second calcification detection model may be a malignant large calcification detection model, and the third calcification detection model may be a malignant small calcification detection model. In other embodiments, N may depend on the classification of the calcification type in the actual application.
The implementation of the apparatus of this embodiment may refer to the implementation of the calcification detection method in breast images, and will not be described herein.
Based on the same technical idea, an embodiment of the present invention provides a computer device, including at least one processor, and at least one memory, wherein the memory stores a computer program, which when executed by the processor, enables the processor to perform the above-mentioned calcification detection method in breast images.
Based on the same technical idea, an embodiment of the present invention provides a computer-readable storage medium, which when executed by a processor within a device, causes the device to perform the above-described calcification detection method in breast images.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, or as a computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (8)
1. A method of detecting calcification in a breast image, comprising:
preprocessing the breast image to obtain a first image;
inputting the first image into N calcification detection models to respectively obtain calcification regions and calcification types of the calcification regions in the first image, wherein N is a positive integer;
based on a preset rule, fusing the calcified region and the calcification type thereof in the obtained first image to obtain a calcification detection result;
the preprocessing the breast image to obtain a first image comprises: binarizing the breast image, and taking the area with the largest area in the binarized breast image as a first image;
the calcification detection model is obtained by respectively training pre-classified training samples of each calcification type, and calcification areas in the training samples are obtained by the following modes:
binarizing the marked calcified region to obtain a circumscribed rectangular frame of the binarized marked calcified region;
taking the area where the circumscribed rectangular frame is located as a calcified area;
the calcification detection model includes: a feature extraction module and a detection frame acquisition module, wherein,
the feature extraction module includes: the system comprises N convolution modules, N maximum pooling layers, N2D deconvolution layers of 2 gamma and tensor superposition layers, wherein the output of the convolution modules is connected with the input of the maximum pooling layers, and the convolution modules are alternately connected with the maximum pooling layers; each convolution module comprises a plurality of convolution blocks, wherein each convolution block comprises: a convolution layer, a batch normalization layer and an activation layer, the activation function being a linear rectification function; the tensor superposition layer is used for outputting tensors of two same dimensions, wherein the tensor of one dimension is output by the pooling layer, the tensor of one dimension is output by the deconvolution layer, and superposition is carried out at the corresponding position to obtain a feature map;
the feature extraction module outputs the feature image output by the tensor superimposed layer and the feature image output by the convolution module to the detection frame acquisition module, the detection frame acquisition module detects a plurality of detection frames in the breast image, reserves the detection frames with the confidence degree larger than a preset threshold value, removes the detection frames with the larger overlapping degree through a non-maximum value inhibition method, and further obtains a final detection result, namely the detection frames belonging to the calcified area are shown in the breast image, and meanwhile, the category of the calcified area is given.
2. The method of claim 1, wherein fusing the calcified region and its calcification type in the obtained first image based on the preset rule to obtain a calcification detection result comprises: the calcification detection result is obtained based on the overlap condition of the first calcified region and the second calcified region and the calcification type.
3. The method of claim 2, wherein the obtaining a calcification detection result based on the overlap condition and the calcification type of the first and second calcified regions comprises:
and determining the intersection ratio of the first calcified region and the second calcified region, and if the intersection ratio is larger than a preset value, merging the first calcified region and the calcified type thereof and the second calcified region and the calcified type thereof based on a merging rule to obtain a calcification detection result.
4. A method as claimed in claim 3, wherein determining the ratio of the intersection of the first calcified region and the second calcified region comprises:
determining the number of first pixel points covered by the intersection of the first calcified region and the second calcified region;
determining the number of second pixel points covered by a union of the first calcified region and the second calcified region;
and determining the ratio of the number of the first pixel points to the number of the second pixel points as the intersection ratio of the first calcified region and the second calcified region.
5. A method as claimed in claim 3, wherein the calcification type comprises: a malignant large calcification type and a malignant small calcification type, the merge rule comprising: combining the malignant large calcification type and the malignant small calcification type into a malignant large calcification type.
6. The method of claim 1, wherein the calcification type comprises: malignant large calcification type, malignant small calcification type, and benign calcification type.
7. A calcification detecting apparatus in a breast image, comprising:
a preprocessing unit for preprocessing the breast image to obtain a first image;
n calcification detection models are used for inputting the first image and respectively outputting calcification areas and calcification types of the calcification areas in the first image, wherein N is a positive integer;
the fusion unit is used for fusing the calcified region and the calcification type thereof in the obtained first image based on a preset rule to obtain a calcification detection result;
the preprocessing the breast image to obtain a first image comprises: binarizing the breast image, and taking the area with the largest area in the binarized breast image as a first image;
the calcification detection model is obtained by respectively training pre-classified training samples of each calcification type, and calcification areas in the training samples are obtained by the following modes:
binarizing the marked calcified region to obtain a circumscribed rectangular frame of the binarized marked calcified region;
taking the area where the circumscribed rectangular frame is located as a calcified area;
the calcification detection model includes: a feature extraction module and a detection frame acquisition module, wherein,
the feature extraction module includes: the system comprises N convolution modules, N maximum pooling layers, N2D deconvolution layers of 2 gamma and tensor superposition layers, wherein the output of the convolution modules is connected with the input of the maximum pooling layers, and the convolution modules are alternately connected with the maximum pooling layers; each convolution module comprises a plurality of convolution blocks, wherein each convolution block comprises: a convolution layer, a batch normalization layer and an activation layer, the activation function being a linear rectification function; the tensor superposition layer is used for outputting tensors of two same dimensions, wherein the tensor of one dimension is output by the pooling layer, the tensor of one dimension is output by the deconvolution layer, and superposition is carried out at the corresponding position to obtain a feature map;
the feature extraction module outputs the feature image output by the tensor superimposed layer and the feature image output by the convolution module to the detection frame acquisition module, the detection frame acquisition module detects a plurality of detection frames in the breast image, reserves the detection frames with the confidence degree larger than a preset threshold value, removes the detection frames with the larger overlapping degree through a non-maximum value inhibition method, and further obtains a final detection result, namely the detection frames belonging to the calcified area are shown in the breast image, and meanwhile, the category of the calcified area is given.
8. A computer device comprising at least one processor, and at least one memory, wherein the memory stores a computer program that, when executed by the processor, enables the processor to perform the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010758673.9A CN111986165B (en) | 2020-07-31 | 2020-07-31 | Calcification detection method and device in breast image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010758673.9A CN111986165B (en) | 2020-07-31 | 2020-07-31 | Calcification detection method and device in breast image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111986165A CN111986165A (en) | 2020-11-24 |
| CN111986165B true CN111986165B (en) | 2024-04-09 |
Family
ID=73444844
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010758673.9A Active CN111986165B (en) | 2020-07-31 | 2020-07-31 | Calcification detection method and device in breast image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111986165B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119579550B (en) * | 2024-11-21 | 2025-07-01 | 昆明医科大学第一附属医院(云南省皮肤病医院) | Intelligent analysis system and method of mammary mammography images based on multi-scale deep learning |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108416360A (en) * | 2018-01-16 | 2018-08-17 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification feature |
| CN109447065A (en) * | 2018-10-16 | 2019-03-08 | 杭州依图医疗技术有限公司 | A kind of method and device of breast image identification |
| CN110390671A (en) * | 2019-07-10 | 2019-10-29 | 杭州依图医疗技术有限公司 | Method and device for detecting breast calcification |
| CN110473186A (en) * | 2019-02-14 | 2019-11-19 | 腾讯科技（深圳）有限公司 | A kind of detection method based on medical image, the method and device of model training |
| CN110647947A (en) * | 2019-09-30 | 2020-01-03 | 杭州依图医疗技术有限公司 | Method and device for lesion fusion |
| CN111428709A (en) * | 2020-03-13 | 2020-07-17 | 平安科技（深圳）有限公司 | Image processing method, image processing device, computer equipment and storage medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108364061B (en) * | 2018-02-13 | 2020-05-05 | 北京旷视科技有限公司 | Arithmetic device, arithmetic execution apparatus, and arithmetic execution method |
| US11010902B2 (en) * | 2018-06-04 | 2021-05-18 | University Of Central Florida Research Foundation, Inc. | Capsules for image analysis |
-
2020
- 2020-07-31 CN CN202010758673.9A patent/CN111986165B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108416360A (en) * | 2018-01-16 | 2018-08-17 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification feature |
| CN109447065A (en) * | 2018-10-16 | 2019-03-08 | 杭州依图医疗技术有限公司 | A kind of method and device of breast image identification |
| CN110473186A (en) * | 2019-02-14 | 2019-11-19 | 腾讯科技（深圳）有限公司 | A kind of detection method based on medical image, the method and device of model training |
| CN110390671A (en) * | 2019-07-10 | 2019-10-29 | 杭州依图医疗技术有限公司 | Method and device for detecting breast calcification |
| CN110647947A (en) * | 2019-09-30 | 2020-01-03 | 杭州依图医疗技术有限公司 | Method and device for lesion fusion |
| CN111428709A (en) * | 2020-03-13 | 2020-07-17 | 平安科技（深圳）有限公司 | Image processing method, image processing device, computer equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| DF-SSD: An Improved SSD Object Detection Algorithm Based on DenseNet and Feature Fusion;SHEPING ZHAI;《IEEE Access》;20200203;全文 * |
| 多尺度遮挡鲁棒的全天候行人检测技术研究;费驰;《中国优秀硕士学位论文全文数据库信息科技辑》;20200215;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111986165A (en) | 2020-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111325739B (en) | Method and device for detecting lung focus and training method of image detection model | |
| CN109447065B (en) | A method and device for breast image recognition | |
| Malarvel et al. | An improved version of Otsu's method for segmentation of weld defects on X-radiography images | |
| CN109363698B (en) | Method and device for identifying mammary gland image signs | |
| Maitra et al. | Technique for preprocessing of digital mammogram | |
| CN110678903B (en) | Systems and methods for analysis of heterotopic ossification in 3D images | |
| JP6267710B2 (en) | System and method for automatically detecting pulmonary nodules in medical images | |
| JP4999163B2 (en) | Image processing method, apparatus, and program | |
| US8401285B1 (en) | Methods, apparatuses, and computer program products for controlling luminance of non-tissue objects within an image | |
| US11783476B2 (en) | System and method for analyzing three-dimensional image data | |
| US8086002B2 (en) | Algorithms for selecting mass density candidates from digital mammograms | |
| CN112365973B (en) | Pulmonary nodule auxiliary diagnosis system based on countermeasure network and fast R-CNN | |
| CN111784701B (en) | Ultrasonic image segmentation method and system combining boundary feature enhancement and multi-scale information | |
| US11684333B2 (en) | Medical image analyzing system and method thereof | |
| EP4022562B1 (en) | Image processing for stroke characterization | |
| CN109363697B (en) | A method and device for identifying breast imaging lesions | |
| Timp et al. | Interval change analysis to improve computer aided detection in mammography | |
| WO2014015424A1 (en) | Tumor classification based on an analysis of a related ultrasonic attenuation map | |
| US11954853B2 (en) | Systems and methods for fast mammography data handling | |
| US20190035075A1 (en) | Method and apparatus for classifying a data point in imaging data | |
| Dobrescu et al. | Diagnosis of breast cancer from mammograms by using fractal measures | |
| JP4651370B2 (en) | Radiation image processing method | |
| CN110738702B (en) | Three-dimensional ultrasonic image processing method, device, equipment and storage medium | |
| CN110060246B (en) | Image processing method, device and storage medium | |
| JP6862147B2 (en) | Image processing device, operation method of image processing device, image processing system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20220318 Address after: Room 705, building 8, No. 1818-2, Wenyi West Road, Yuhang District, Haidian District, Beijing 100089 Applicant after: BEIJING SHENRUI BOLIAN TECHNOLOGY Co.,Ltd. Applicant after: SHENZHEN DEEPWISE BOLIAN TECHNOLOGY Co.,Ltd. Address before: 200051 units 06 and 07, 23 / F, 523 Loushanguan Road, Changning District, Shanghai Applicant before: SHANGHAI YIZHI MEDICAL TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |