CN116030459A - Detection method, device and storage medium for recognizing malaria parasites - Google Patents

Abstract

The application relates to a detection method, device and storage medium for identifying malaria parasites. The method includes: acquiring a blood smear image to be detected; inputting the blood smear image to be detected into a malaria parasite identification model, and using the malaria parasite identification model to detect the type of malaria parasite and its developmental stage; acquiring the malaria parasite identification The detection results output by the model. The solution provided by this application can automatically detect the morphological characteristics of different species of Plasmodium and different life cycles of Plasmodium in the image through the Plasmodium recognition model, so as to realize the automatic classification and identification of various Plasmodium.

Description

Detection method, device and storage medium for identifying malaria parasites

technical field

The present application relates to the technical field of detection, in particular to a detection method, device and storage medium for identifying malaria parasites.

Background technique

Malaria is a serious infectious disease caused by Plasmodium parasites in peripheral blood. In related technologies, finding Plasmodium in blood smears by manual microscopy is the "gold standard" for malaria diagnosis.

Manual microscopy is easy to operate, low in cost, and can identify species of parasites. However, when the detection volume is large, it often leads to missed or false detections of Plasmodium due to lack of experience, fatigue, and inattention of technicians.

Contents of the invention

In order to solve or partially solve the problems existing in related technologies, this application provides a detection method, device and storage medium for identifying malaria parasites, which can realize automatic classification and detection of various malaria parasites, and improve the shortcomings of manual microscopy.

The first aspect of the present application provides a detection method for identifying Plasmodium, comprising:

Obtain the image of the blood smear to be detected;

Input the image of the blood smear to be detected into the malaria parasite recognition model, and the malaria parasite species and development stage are detected by the malaria parasite recognition model;

Obtain the detection results output by the Plasmodium identification model.

In one embodiment, the Plasmodium recognition model is trained in the following manner:

Collecting a Plasmodium blood smear image, marking the type of Plasmodium corresponding to the Plasmodium blood smear image and its developmental stage;

A convolutional neural network model is used to learn and train the Plasmodium blood smear image and the Plasmodium species and developmental stages corresponding to the Plasmodium blood smear image to obtain a Plasmodium identification model.

In one embodiment, the acquisition of the image of the blood smear to be detected includes:

Scanning the blood smear to be detected to obtain an image of the blood smear to be detected; wherein, the blood smear to be detected is a thick blood film blood smear to be detected;

The collection of Plasmodium blood smear images includes:

The Plasmodium blood smear is scanned to obtain a Plasmodium blood smear image; wherein, the Plasmodium blood smear is a Plasmodium thick blood film blood smear.

In one embodiment, the convolutional neural network model is used to learn and train the Plasmodium blood smear image and the Plasmodium species and developmental stages corresponding to the Plasmodium blood smear image, to obtain Plasmodium identification models, including:

Carrying out image feature recognition and extraction on the Plasmodium blood smear image, and matching with the Plasmodium species corresponding to the Plasmodium blood smear image and the label of its developmental stage to obtain a training set;

A convolutional neural network model is used to learn and train the training set to obtain a malaria parasite identification model.

In one embodiment, the training set is learned and trained using a convolutional neural network model to obtain a malaria parasite identification model, including:

Using Faster-RCNN to learn and train the training set to obtain a malaria parasite identification model.

In one embodiment, the input of the image of the blood smear to be detected into the Plasmodium identification model, and the detection of the type of Plasmodium and the developmental stage of the Plasmodium by the Plasmodium identification model include:

According to the morphological characteristics of the malaria parasites in the blood smear, the blood smear image to be detected is detected for the species of the malaria parasite and its developmental stage; wherein, the morphological characteristics of the malaria parasites in the blood smear include Plasmodium The shape, structure, size and staining of protozoa.

In one embodiment, the staining of the malaria parasite includes the staining of the nucleus, cytoplasm and malaria pigment of the malaria parasite; or,

The staining of the malaria parasite includes the staining of the nucleus and cytoplasm of the malaria parasite.

The second aspect of the present application provides a detection device for identifying Plasmodium, comprising:

An image acquisition module, configured to acquire an image of a blood smear to be detected;

A detection module, configured to input the image of the blood smear to be detected obtained by the image acquisition module into the malaria parasite identification model, and the malaria parasite identification model is used to detect the type of malaria parasite and its developmental stage;

The result output module is used to obtain the detection result output by the malaria parasite identification model.

The third aspect of the present application provides an electronic device, including:

processor; and

A memory, on which executable codes are stored, which, when executed by the processor, cause the processor to perform the method as described above.

A fourth aspect of the present application provides a computer-readable storage medium, on which executable code is stored, and when the executable code is executed by a processor of an electronic device, the processor is caused to execute the above-mentioned method.

The technical solution provided by the present application may include the following beneficial effects: through the Plasmodium recognition model, the Plasmodium species and developmental stages can be identified and classified on Plasmodium blood smear images, thereby realizing automatic detection of various types of Plasmodium, reducing Rapid diagnosis of small malaria relies on skilled professionals to achieve efficient and accurate detection of malaria parasites.

The technical solution of the present application can also: use the Plasmodium identification model to identify and classify the Plasmodium species and developmental stages of Plasmodium in thick blood film blood smears, with a wide range of observation and a high detection rate of Plasmodium , and does not need to identify and calibrate red blood cells first, it can directly identify the morphological characteristics of Plasmodium and improve the detection efficiency of various Plasmodium.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.

Description of drawings

The above and other objects, features and advantages of the present application will become more apparent by describing the exemplary embodiments of the present application in more detail with reference to the accompanying drawings, wherein, in the exemplary embodiments of the present application, the same reference numerals generally represent same parts.

Fig. 1 is a schematic flow chart of the detection method for identifying Plasmodium shown in the embodiment of the present application;

Fig. 2 is a schematic flow chart of the training method of the Plasmodium recognition model shown in the embodiment of the present application;

Fig. 3 is the schematic diagram of making thick blood film and thin blood film smear shown in the embodiment of the present application;

FIG. 4 is a schematic diagram of the architecture of the Faster-RCNN shown in the embodiment of the present application;

Fig. 5 is a schematic structural diagram of a detection device for identifying malaria parasites shown in an embodiment of the present application;

FIG. 6 is a schematic structural diagram of an electronic device shown in an embodiment of the present application.

Detailed ways

Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although embodiments of the present application are shown in the drawings, it should be understood that the present application may be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of this application to those skilled in the art.

The terminology used in this application is for the purpose of describing particular embodiments only, and is not intended to limit the application. As used in this application and the appended claims, the singular forms "a", "the", and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It should also be understood that the term "and/or" as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

It should be understood that although the terms "first", "second", "third" and so on may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of the present application, first information may also be called second information, and similarly, second information may also be called first information. Thus, a feature defined as "first" and "second" may explicitly or implicitly include one or more of these features. In the description of the present application, "plurality" means two or more, unless otherwise specifically defined.

In related technologies, when using manual microscopic examination to diagnose malaria, technicians need to undergo professional training, accumulate long-term experience, possess proficient skills, and also need continuous training to maintain technical capabilities. This method is not a reliable means of screening when manual microscopic examination is performed by inexperienced technicians, especially in economically underdeveloped areas with relatively high malaria prevalence. On the other hand, manually moving the focus during manual microscope inspection and using the naked eye to observe and interpret the results under the microscope is labor-intensive and low-efficiency. Long-term work under the microscope can also cause eye pain, vision loss, and dizziness for technicians. Other consequences may affect the detection results, resulting in missed or false detection of malaria parasites.

In view of the above problems, the embodiment of the present application provides a detection method for identifying malaria parasites, which can realize automatic classification and detection of various malaria parasites, and improve the shortcomings of manual microscopy.

The technical solutions of the embodiments of the present application are described in detail below with reference to the accompanying drawings.

Fig. 1 is a schematic flowchart of a detection method for identifying Plasmodium shown in an embodiment of the present application.

Referring to Figure 1, detection methods to identify Plasmodium include:

S110. Obtain an image of the blood smear to be detected.

This step may be to scan the blood smear to be detected to obtain an image of the blood smear to be detected. Wherein, the scanning process of the blood smear to be tested adopts an automatic microscopic scanning system, and the automatic microscopic scanning system includes an automatic scanner, a microscope, a camera, a computer, and the like.

Wherein, scanning the blood smear to be detected may be scanning the thick blood film blood smear to be detected to obtain an image of the blood smear to be detected.

S120. Input the image of the blood smear to be detected into the malaria parasite identification model, and the malaria parasite identification model detects the type and developmental stage of the malaria parasite.

Among them, the malaria parasite recognition model can be trained as follows:

S121. Collect the Plasmodium blood smear image, and label the Plasmodium species corresponding to the Plasmodium blood smear image and its developmental stage.

Wherein, collecting the Plasmodium blood smear image may be scanning the Plasmodium blood smear image to obtain the Plasmodium blood smear image; the Plasmodium blood smear may be a Plasmodium thick blood film blood smear.

S122. Using a convolutional neural network model to learn and train the Plasmodium blood smear image and the Plasmodium species and developmental stage labels corresponding to the Plasmodium blood smear image, to obtain a Plasmodium identification model.

This step can be performed on the image feature recognition and extraction of the Plasmodium blood smear image first, and matching the image feature with the label of the Plasmodium species and developmental stage to obtain a training set, and then using the convolutional neural network model to train The set is learned and trained to obtain a malaria parasite recognition model.

Wherein, in the image feature recognition and extraction process, the image feature recognition and extraction may be performed on the Plasmodium blood smear image according to the morphological characteristics of the Plasmodium in the blood smear. The morphological characteristics of Plasmodium in blood smears include the shape, structure, size, and staining of Plasmodium; among them, the staining of Plasmodium can be composed of three staining conditions: nucleus, cytoplasm and malaria pigment of Plasmodium; or, Plasmodium The staining of the parasite can be composed of the nucleus and cytoplasm of the malaria parasite.

Among them, learning and training can be carried out by using Faster-RCNN to obtain a malaria parasite recognition model.

S130. Obtain the detection result output by the malaria parasite identification model.

The Plasmodium identification model can identify and judge whether there is Plasmodium, the type of Plasmodium and its developmental stage in the blood smear to be detected, and the output detection results include whether there is Plasmodium, the type of Plasmodium and its developmental stage. Among them, the species of Plasmodium include Plasmodium vivax, Plasmodium falciparum, Plasmodium malariae and Plasmodium ovale; the developmental stages of Plasmodium refer to the developmental stages of the above four Plasmodium respectively, including small trophozoites ( Ring body), macrotrophozoite, schizont, gametophyte and other stages.

The technical solution of the present application first obtains the image of the blood smear to be detected by scanning, and then uses the Plasmodium recognition model to classify and detect the image of the blood smear to be detected, so as to realize the automatic detection of the Plasmodium under the microscope, achieving high efficiency and accuracy The effect of identifying the species of Plasmodium and its developmental stage, reducing the dependence of rapid malaria diagnosis on skilled professionals, and improving the shortcomings of manual microscopy.

The scheme of this application is further described in detail below.

The Plasmodium identification model in the scheme of this application includes firstly using the collected Plasmodium blood smear images to obtain the training set, and then using the convolutional neural network model to learn and train the training set, thereby obtaining the Plasmodium identification model, and then using the Plasmodium The protozoa identification model is used to detect the blood smear to be detected.

Fig. 2 is a schematic flowchart of a method for training a Plasmodium identification model shown in an embodiment of the present application.

Referring to Figure 2, the training methods of the malaria parasite identification model include:

S121. Collect the Plasmodium blood smear image, and label the Plasmodium species corresponding to the Plasmodium blood smear image and its developmental stage.

This step may be scanning the Plasmodium blood smear to obtain an image of the Plasmodium blood smear. Among them, the scanning process of the Plasmodium blood smear adopts an automatic microscopic scanning system, which includes an automatic scanner, a microscope, a camera, a computer, and the like.

When the automatic micro-scanning system scans the blood smear sample, each field of view will perform a subtle focusing operation to collect a clear image of the field of view. During the focusing process, the automatic micro-scanning system will use Energy (energy ) algorithm to calculate the sharpness of the image, select the image with the highest value as the representative image of the field of view, and obtain the Plasmodium blood smear image. In the specific operation process, it is necessary to classify the known species of Plasmodium, scan the blood smear of the same Plasmodium under the oil microscope, and set enough fields of view for each blood smear, such as setting to scan 1000 fields of view, Obtain enough image information and save relevant image information.

A blood smear refers to a smear made by applying blood to a glass slide. The blood smears used for the detection of Plasmodium by microscope include thick blood film blood film and thin blood film film film. Fig. 3 is a schematic diagram of making thick blood film and thin blood film blood smear shown in the embodiment of the present application, in which Sample1 is a thick blood film smear, and Sample2 is a thin blood film blood smear. See Figure 3, Sample 1 thick blood film is to take 4μL ~ 5μL blood drop and smear on the glass slide, draw a circle from the inside to the outside to form a circular thick blood film with a diameter of 0.8cm ~ 1.0cm, and the thickness is within the field of view of 1 oil immersion It is advisable to see 5 to 10 white blood cells; Sample2 thin blood film is to take 1 μL to 1.5 μL of blood and spread it on a glass slide to form a tongue-shaped thin blood film with a width of 2 cm. After the blood smear is prepared, it is necessary to hemolyze the thick blood film, fix the thin blood film with methanol, and then perform Giemsa staining on the thick blood film and the thin blood film respectively. Add 1 drop of cedar oil or special immersion oil to the stained and dried blood film, and check the blood film with a 100× oil immersion objective lens and an optical microscope with 5× or 10× eyepieces.

Wherein, the collected Plasmodium blood smear image may be a Plasmodium thick blood film blood smear. A thick blood film requires a large amount of blood, a large field of view can be observed, and a high detection rate of malaria parasites.

The labeling of the Plasmodium species and developmental stages corresponding to the Plasmodium blood smear images may be obtained by interpreting and labeling the Plasmodium species and developmental stages after the Plasmodium blood smear images are collected.

Wherein, the annotation can be obtained by manual interpretation and annotation. Manual interpretation and labeling refers to organizing a number of professional technicians who have received professional technical training and many years of experience in malaria microscopy to interpret the scanned images under the microscope one by one, and mark the type and developmental stage of the malaria parasite in each image. The types of Plasmodium include Plasmodium vivax, Plasmodium falciparum, Plasmodium malariae and Plasmodium ovale; the developmental stages of Plasmodium refer to the developmental stages of the above four Plasmodium respectively, including small trophozoites (circular body), macrotrophoblast, schizont, gametophyte and other stages. In addition, when performing manual interpretation and labeling, platelets, white blood cells, and other stained impurities can also be interpreted and marked.

Under the microscope, manual interpretation can identify the type of Plasmodium and its developmental stage according to the shape, structure, size, and staining of the Plasmodium. See Table 1 for details.

Table 1 Morphological identification of thick blood film of four kinds of Plasmodium (Gigi's staining)

S122. Perform image feature recognition and extraction on the Plasmodium blood smear image, and perform matching with the corresponding labels on the Plasmodium blood smear image to obtain a training set.

Image feature recognition and extraction may include the following steps: image segmentation, feature extraction. Wherein, in the image segmentation step, it is first necessary to locate the stained object by the color intensity value, and then divide the stained region into various elements, such as white blood cells, platelets, malaria parasites, etc., this step can be normalized cutting ( Normalized cut, NCut) method. After the blood smear is stained with Giemsa, red blood cells are light red, eosinophils are bright red, neutrophils are purple blue, lymphocytes and Plasmodium cytoplasm are blue or light blue, and Plasmodium nuclei are Purplish red, malarin is brownish yellow, tan or dark brown.

In the thin blood film, cells are stained after methanol fixation, which preserves the complete shape of red blood cells. Plasmodium exists in the cytoplasm of infected red blood cells, so the computer algorithm can first identify red blood cells through calibration, and then identify whether the Plasmodium is detected by the presence of Plasmodium. However, because the thick blood film is first hemolyzed and then stained, the red blood cells are destroyed, and the malaria parasites observed under the microscope are outside the cells, so the computer algorithm cannot track and judge the malaria parasites through the red blood cells. In the feature extraction step, feature extraction is performed on the image according to features such as the shape, structure, size, and staining of the malaria parasite, so as to identify the malaria parasite.

Among them, the shape of Plasmodium includes "!" shape, "." shape, "V" shape, bird shape, bird's eye shape, ring shape, broken ring shape, round shape, oval shape, crescent shape, sausage shape etc.; the structure of Plasmodium includes the nucleus located in the cytoplasm, the nucleus located outside the cytoplasm, merozoites, and gametophytes; The staining of Plasmodium in the blood smear can be composed of three staining conditions of nucleus, cytoplasm and malaria pigment; or, when the malaria parasite has not yet produced malaria pigment (malar pigment can be seen in all stages except rings) When , it can be composed of two staining conditions of the nucleus and cytoplasm of Plasmodium. For example, after Giemsa staining, the nucleus of Plasmodium is purple, the cytoplasm is blue or light blue, and the malaria pigment is brown, brown or dark brown.

In addition, features such as malaria pigment and parasitized red blood cells of Plasmodium can also be extracted. The characteristics of malaria parasite include the presence or absence of malaria pigment, the color of malaria pigment, the particle size of malaria pigment, and the distribution of malaria pigment.

According to the morphological characteristics of the Plasmodium blood smear under the microscope, after the image recognition and extraction of the Plasmodium blood smear, the image features can also be matched with the annotations of the type and developmental stage of the Plasmodium in S121, and the labeled The image features of the data, i.e. the training set. During the matching process, the Python-based tool software LabelImg can be used to label the identified and extracted Plasmodium blood smear images to obtain image features with labeled data. The labeled content includes the species of Plasmodium and its developmental stage, Platelets, white blood cells, other staining impurities, etc.

S123. Using Faster-RCNN to learn and train the training set to obtain a malaria parasite identification model.

Among them, Faster-RCNN, as a target detection system, includes three modules, the first module is CNN (Convolutional Neural Networks, Convolutional Neural Networks) for extracting feature regions, and the second module is used for training the first One module extracts the feature region, generates the RPN (Region Proposal Network, candidate region generation network) of the candidate region, and the third module uses RoIPooling (region of interest pooling) technology to match the feature region with the candidate region, and then proceeds Classification and regression, so as to realize the construction of malaria parasite identification model. Fig. 4 is a schematic diagram of the architecture of Faster-RCNN shown in the embodiment of the present application. Combined with Figure 4, Faster-RCNN's learning and training of the training set includes the following three steps:

1) Generate a feature map:

As a commonly used target detection model based on CNN (Convolutional Neural Networks, convolutional neural network), Faster-RCNN uses 13 convolutional layers, 13 ReLU (Rectified Linear Unit, corrected linear unit) activation function layers and 4 The pooling layer extracts feature maps from the input Plasmodium blood smear image, and the process of extracting feature maps is carried out according to the shape, structure, size, and staining of Plasmodium.

2) Generate candidate regions:

The feature map input to RPN (Region Proposal Network, candidate area generation network) generates "anchor frames" of different sizes after RPN, and then uses the Softmax function (normalization function) to distinguish these "anchor frame" areas into foreground ( target) and background (non-target), while the "bounding box regression" algorithm corrects these "anchor boxes" into more accurate "candidate regions". Among them, the foreground (target object) refers to Plasmodium, and the background (non-target object) refers to white blood cells, platelets, and other impurities other than Plasmodium.

3) Classify the candidate area and return the coordinates and credibility of the target object:

First, use RoIPooling (region of interest pooling) technology to match the feature map in S241 with the candidate area in S242 to obtain the feature data of the corresponding area. The fully connected layer uses these feature data to classify and locate the target object, and returns the target object Corresponding coordinates (the corresponding position coordinates of the image taken in the thick blood film recorded when scanning the Plasmodium thick blood film blood smear) and reliability. Among them, the classification of the target includes the classification of the species of Plasmodium and the classification of the developmental stages of the various types of Plasmodium. The classification of the target uses the Softmax-Loss function (by softmax (normalization function) and cross-entropy The loss function (cross entropy loss function) is realized; the positioning of the target refers to the corresponding coordinates of the malaria parasite in the corresponding thick blood film smear, and the positioning of the target adopts the "bounding box regression" combined with the Smooth L1 Loss function "Algorithm implementation.

When using the above-mentioned Plasmodium identification model to detect the blood smear to be detected, the Plasmodium identification model can directly detect the type and development stage of Plasmodium, thereby realizing the automatic classification and detection of Plasmodium.

Referring to FIG. 1 , in the detection method of malaria parasites, it is first necessary to obtain an image of the blood smear to be detected. Obtaining the image of the blood smear to be detected may be scanning the thick blood film blood smear to be detected, wherein the scanning process of the Plasmodium blood smear adopts an automatic micro-scanning system, and the automatic micro-scanning system includes an automatic scanner, a microscope, Cameras, computers, etc. When scanning, the position coordinates of the captured image corresponding to the thick blood film can be recorded, and the questionable field of view can be returned to this coordinate, and the field of view will be confirmed by professional technicians, and the confirmation result will be returned to the malaria parasite identification model , supplemented to the Plasmodium recognition model.

In the detection method of Plasmodium, secondly, the obtained blood smear to be detected needs to be detected by using the above-mentioned Plasmodium identification model. During the detection process, the malaria parasite identification model can detect the malaria parasites in the blood smear image to be detected according to the morphological characteristics of the malaria parasites in the thick blood smear. The morphological characteristics of Plasmodium in blood smears include the shape, structure, size, and staining of Plasmodium. In addition, auxiliary detection and judgment can be performed based on the characteristics of Plasmodium and parasitized red blood cells.

Under the microscope, there are no obvious red blood cells in the thick blood film, and the volume of white blood cells (including neutrophils, eosinophils, lymphocytes, etc.) is several times larger than that of malaria parasites, so they can be screened out by size leukocyte. The size of platelets is close to that of Plasmodium, but platelets do not have a nucleus structure, so platelets can be screened by the nucleus structure. After the thick blood film blood smear is stained with Giemsa, eosinophils are bright red, neutrophils are purple-blue, lymphocytes and Plasmodium cytoplasm are blue or light blue, and Plasmodium nuclei are purple-red , malarin is brownish yellow, brown or dark brown. Plasmodium can be detected and judged through the morphological characteristics such as the size, nuclear structure, and staining of Plasmodium in thick blood film smears, that is, whether there are Plasmodium in thick blood film blood smears can be detected. Then according to the shape, structure, size, malaria pigment, etc. of the four different types of malaria parasites in Table 1 and the characteristics of the parasited red blood cells, the classification of malaria parasites can be judged, for example: whether there is malaria pigment, the color of malaria pigment, the malaria pigment The size of the particles, the distribution of malaria pigment, etc., the parasited red blood cells include whether the red blood cell "shadow", Scherner's point, Mao's point, etc. are visible, and the type of malaria parasite and its developmental stage in the thick blood film blood smear are detected , so as to realize the automatic detection and classification of malaria parasites.

The technical solution of the present application is to collect images of Plasmodium thick blood film blood smears, then identify and extract image features of Plasmodium thick blood film images, and learn and train by the convolutional neural network model, thereby Obtain a Plasmodium recognition model. The Plasmodium identification model can detect and identify the Plasmodium under the thick blood film, so as to realize the automatic classification and recognition of the Plasmodium under the microscope image, and achieve the effect of efficient and accurate identification of the Plasmodium species and their developmental stages. Reduce reliance on skilled professionals for rapid malaria diagnosis. On the other hand, compared with the identification of Plasmodium through thin blood film, thick blood film blood smear requires more blood, has a larger field of view, has a higher detection rate than thin blood film, and takes less time for scanning and photographing, which can shorten the detection time. time, improve the detection efficiency; and do not need to identify and calibrate the red blood cells first, can directly identify the shape of the malaria parasite, directly classify the malaria parasite, and complete the detection of the presence or absence of the malaria parasite, the determination of the type of the malaria parasite and the developmental stage of the malaria parasite, The detection steps are simplified, the detection time is further shortened, and the detection efficiency is improved.

Corresponding to the foregoing embodiment of the method for implementing application functions, the present application also provides a detection device for identifying malaria parasites, electronic equipment, and corresponding embodiments.

Fig. 5 is a schematic structural diagram of a detection device for identifying malaria parasites shown in an embodiment of the present application.

Referring to FIG. 5 , a detection device 500 for identifying malaria parasites includes an image acquisition module 510 , a detection module 520 , and a result output module 530 .

An image acquisition module 510, configured to acquire an image of the blood smear to be detected.

Wherein, the image acquisition module 510 can be used to scan the blood smear to be detected, and acquire the image of the blood smear to be detected; wherein, the blood smear to be detected can be a thick blood film smear to be detected.

The detection module 520 is configured to input the image of the blood smear to be detected acquired by the image acquisition module 510 into the Plasmodium identification model, and the Plasmodium identification model detects the type and development stage of the Plasmodium.

Among them, the Plasmodium identification model can detect the type of Plasmodium and its developmental stage according to the morphological characteristics of Plasmodium in the blood smear to be detected; wherein, the morphological characteristics of Plasmodium in the blood smear include Plasmodium The shape, structure, size, and staining of the parasite; the staining of the malaria parasite includes the staining of the nucleus, cytoplasm, and malarin of the malaria parasite, or the staining of the malaria parasite includes the nucleus and cytoplasm of the malaria parasite.

The result output module 330 is configured to obtain the detection result output by the malaria parasite identification model.

According to the technical solution of the present application, the image of the blood smear to be detected is obtained by scanning first, and then the image of the blood smear to be detected is detected by the Plasmodium recognition model, so as to realize the automatic classification and identification of the Plasmodium under the microscope, achieving high efficiency and accuracy The effect of identifying the species of Plasmodium and its developmental stage reduces the reliance on skilled professionals for the rapid diagnosis of malaria.

Regarding the apparatus in the above embodiments, the specific manner in which each module executes operations has been described in detail in the embodiments related to the method, and will not be described in detail here.

FIG. 6 is a schematic structural diagram of an electronic device shown in an embodiment of the present application.

Referring to FIG. 6 , an electronic device 600 includes a memory 610 and a processor 620 .

The processor 620 can be a central processing unit (Central Processing Unit, CPU), and can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), on-site Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like.

The memory 610 may include various types of storage units such as system memory, read only memory (ROM), and persistent storage. Wherein, the ROM may store static data or instructions required by the processor 620 or other modules of the computer. The persistent storage device may be a readable and writable storage device. Persistent storage may be a non-volatile storage device that does not lose stored instructions and data even if the computer is powered off. In some embodiments, the permanent storage device adopts a large-capacity storage device (such as a magnetic or optical disk, flash memory) as the permanent storage device. In some other implementations, the permanent storage device may be a removable storage device (such as a floppy disk, an optical drive). The system memory can be a readable and writable storage device or a volatile readable and writable storage device, such as dynamic random access memory. System memory can store some or all of the instructions and data that the processor needs at runtime. In addition, the memory 610 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (such as DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and magnetic disks and/or optical disks may also be used. In some embodiments, memory 610 may include a readable and/or writable removable storage device, such as a compact disc (CD), a read-only digital versatile disc (e.g., DVD-ROM, dual-layer DVD-ROM), Read-only Blu-ray Disc, Super Density Disc, Flash memory card (such as SD card, min SD card, Micro-SD card, etc.), magnetic floppy disk, etc. Computer-readable storage media do not contain carrier waves and transient electronic signals transmitted by wireless or wire.

Executable codes are stored in the memory 610 , and when the executable codes are processed by the processor 620 , the processor 620 may execute part or all of the methods mentioned above.

In addition, the method according to the present application can also be implemented as a computer program or computer program product, which includes computer program code instructions for executing some or all of the steps in the above-mentioned method of the present application.

Alternatively, the present application may also be implemented as a computer-readable storage medium (or a non-transitory machine-readable storage medium or a machine-readable storage medium), on which executable code (or computer program or computer instruction code) is stored, When the executable code (or computer program or computer instruction code) is executed by the processor of the electronic device (or server, etc.), the processor is made to perform part or all of the steps of the above-mentioned method according to the present application.

Having described various embodiments of the present application above, the foregoing description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and alterations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principle of each embodiment, practical application or improvement of technology in the market, or to enable other ordinary skilled in the art to understand each embodiment disclosed herein.

Claims (10)

1. A detection method for identifying malaria parasites, comprising: Obtain the image of the blood smear to be detected; Input the image of the blood smear to be detected into the malaria parasite recognition model, and the malaria parasite species and development stage are detected by the malaria parasite recognition model; Obtain the detection results output by the Plasmodium identification model. 2. The detection method according to claim 1, characterized in that: the malaria parasite identification model is trained and obtained in the following manner: Collecting a Plasmodium blood smear image, marking the type of Plasmodium corresponding to the Plasmodium blood smear image and its developmental stage; A convolutional neural network model is used to learn and train the Plasmodium blood smear image and the Plasmodium species and developmental stages corresponding to the Plasmodium blood smear image to obtain a Plasmodium identification model. 3. The detection method according to claim 2, characterized in that: The acquisition of the image of the blood smear to be detected includes: Scanning the blood smear to be detected to obtain an image of the blood smear to be detected; wherein, the blood smear to be detected is a thick blood film blood smear to be detected; The collection of Plasmodium blood smear images includes: The Plasmodium blood smear is scanned to obtain a Plasmodium blood smear image; wherein, the Plasmodium blood smear is a Plasmodium thick blood film blood smear. 4. The detection method according to claim 2, characterized in that: said Plasmodium blood smear image and the Plasmodium species corresponding to the Plasmodium blood smear image and their location The labeling of the developmental stage is used for learning and training, and the identification model of malaria parasites is obtained, including: Carrying out image feature recognition and extraction on the Plasmodium blood smear image, and matching with the Plasmodium species corresponding to the Plasmodium blood smear image and the label of its developmental stage to obtain a training set; A convolutional neural network model is used to learn and train the training set to obtain a malaria parasite identification model. 5. detection method according to claim 4, is characterized in that: described training set is studied and trained by adopting convolutional neural network model, obtains malaria parasite identification model, comprises: Using Faster-RCNN to learn and train the training set to obtain a malaria parasite identification model. 6. The detection method according to any one of claims 1-5, characterized in that: the blood smear image to be detected is input into a Plasmodium recognition model, and the Plasmodium species and the detected Plasmodium are determined by the Plasmodium recognition model. Detection at developmental stages, including: According to the morphological characteristics of the malaria parasites in the blood smear, the blood smear image to be detected is detected for the species of the malaria parasite and its developmental stage; wherein, the morphological characteristics of the malaria parasites in the blood smear include Plasmodium The shape, structure, size and staining of protozoa. 7. The detection method according to claim 6, characterized in that: the staining situation of the malaria parasite comprises the staining situation of the nucleus of the malaria parasite, cytoplasm and malaria pigment; or, The staining of the malaria parasite includes the staining of the nucleus and cytoplasm of the malaria parasite. 8. A detection device for identifying malaria parasites, comprising: An image acquisition module, configured to acquire an image of a blood smear to be detected; A detection module, configured to input the image of the blood smear to be detected obtained by the image acquisition module into the malaria parasite identification model, and the malaria parasite identification model is used to detect the type of malaria parasite and its developmental stage; The result output module is used to obtain the detection result output by the malaria parasite identification model. 9. An electronic device, characterized in that it comprises: processor; and A memory on which executable code is stored, and when the executable code is executed by the processor, causes the processor to execute the method according to any one of claims 1-7. 10. A computer-readable storage medium, characterized in that executable code is stored thereon, and when the executable code is executed by a processor of an electronic device, the processor executes the any one of the methods described.

Images