A chest radiograph screening classification prediction method and device, electronic equipment and storage medium
By enhancing the contrast of chest X-ray images and classifying them using deep learning, the problems of misdiagnosis and missed diagnosis in chest X-ray diagnosis have been solved, achieving efficient automated diagnosis and interpretable results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- PEKING UNIVERSITY FIRST HOSPITAL (PEKING UNIVERSITY FIRST CLINICAL MEDICAL COLLEGE)
- Filing Date
- 2022-11-21
- Publication Date
- 2026-06-05
AI Technical Summary
In current technology, the diagnosis of chest X-rays relies on manual identification, which is prone to misdiagnosis and missed diagnosis, and is inefficient.
A deep learning algorithm is used to enhance the contrast of chest X-ray images, and a pre-trained classification model is used for automatic classification to generate an interpretable model activation diagram.
It improves diagnostic efficiency, reduces misdiagnosis and missed diagnosis, and provides the possibility of quality control for telemedicine and health check-up centers.
Smart Images
Figure CN116129170B_ABST
Abstract
Description
Technical Field
[0001] One or more embodiments of the present invention relate to the field of medical diagnostic technology, and in particular to a chest X-ray screening classification prediction method, device, electronic device and storage medium. Background Technology
[0002] Chest X-ray (CXR) is a first-line imaging method for diagnosing respiratory diseases and accounts for a large proportion of the daily work in medical radiology departments. Clinical scenarios using CXR include outpatient examinations, routine preoperative examinations, health checkups, and emergency examinations. Currently, in the first three cases, CXR results are mostly normal; the purpose of imaging examinations or screenings is to detect the few abnormal cases for further treatment.
[0003] However, in current clinical practice, the above methods are used for manual differential diagnosis by clinicians. However, due to the different experience of each clinician, misdiagnosis and missed diagnosis are easy to occur, and the efficiency of manual differential diagnosis is low. Summary of the Invention
[0004] To address the problems existing in the prior art, this invention provides a chest X-ray screening classification prediction method, device, electronic device, and storage medium. Based on a deep learning algorithm, it uses routine chest X-rays to perform interpretable classification of "no findings" and "finds found" in outpatient, emergency, preoperative routine, and physical examination populations, thereby improving the work efficiency of clinicians, reducing missed diagnoses and misdiagnoses, and also providing possibilities for quality control in telemedicine or physical examination centers.
[0005] To achieve the above objectives, the present invention provides the following technical solution:
[0006] In a first aspect, the present invention provides a chest X-ray screening classification prediction method, the method comprising the following steps:
[0007] Acquire the chest X-ray image to be screened, and perform contrast enhancement processing on the chest X-ray image to obtain a contrast-enhanced image;
[0008] The contrast-enhanced image is input into a pre-trained classification model to determine the category and corresponding score of the chest X-ray image, as well as the corresponding interpretability model activation diagram.
[0009] In one possible implementation, performing contrast enhancement processing on the chest X-ray image to obtain a contrast-enhanced image includes the following steps:
[0010] Estimate the histogram distribution of the chest X-ray image and calculate the cumulative distribution function value;
[0011] The boundary value of the contrast stretching of the chest X-ray image is estimated based on the cumulative distribution function value, and the contrast stretching of the chest X-ray image is performed using the boundary value to determine the contrast stretched image.
[0012] The radius of the disc-shaped morphological structural unit is determined based on the size of the chest X-ray image;
[0013] Based on the radius of the disk-shaped morphological structural unit, morphological top-hat filtering and morphological bottom-hat filtering are performed on the contrast-stretched image to determine the morphological top-hat filtered image and the morphological bottom-hat filtered image.
[0014] A contrast-enhanced image is determined based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image.
[0015] In one possible implementation, the radius of the disk-shaped morphological structural unit is determined according to the following formula:
[0016] R = Max(sx, xy) * 3%
[0017] Where sx and sy represent the length and width of the chest X-ray image, respectively, and Max represents the maximum value of the length and width.
[0018] In one possible implementation, the contrast-enhanced image is determined based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image using the following formula:
[0019] Contrast-enhanced chest X-ray image = contrast-stretched image + morphological top-hat filtered image - morphological bottom-hat filtered image.
[0020] In one possible implementation, the classification model consists of an image feature extraction module, an activation map generation module, and a classification module, wherein,
[0021] The image feature extraction module is used to extract image features from the input contrast-enhanced image;
[0022] The activation map generation module is used to generate an interpretable model activation diagram based on the image features;
[0023] The classification module is used to determine the category and corresponding score of the chest X-ray image based on the image features.
[0024] Secondly, the present invention provides a chest X-ray screening and classification device, characterized in that the device comprises:
[0025] The contrast intensity processing module is used to perform contrast enhancement processing on the acquired chest X-ray images to be screened to obtain contrast-enhanced images;
[0026] The screening classification prediction module is used to input the contrast-enhanced image into a pre-trained classification model to determine the category and corresponding score of the chest X-ray image, as well as the corresponding interpretability model activation diagram.
[0027] In one possible implementation, the contrast intensity processing module is specifically used for:
[0028] Estimate the histogram distribution of the chest X-ray image and calculate the cumulative distribution function value;
[0029] The boundary value of the contrast stretching of the chest X-ray image is estimated based on the cumulative distribution function value, and the contrast stretching of the chest X-ray image is performed using the boundary value to determine the contrast stretched image.
[0030] The radius of the disc-shaped morphological structural unit is determined based on the size of the chest X-ray image;
[0031] Based on the radius of the disk-shaped morphological structural unit, morphological top-hat filtering and morphological bottom-hat filtering are performed on the contrast-stretched image to determine the morphological top-hat filtered image and the morphological bottom-hat filtered image.
[0032] A contrast-enhanced image is determined based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image.
[0033] In one possible implementation, the classification model consists of an image feature extraction module, an activation map generation module, and a classification module, wherein,
[0034] The image feature extraction module is used to extract image features from the input contrast-enhanced image;
[0035] The activation map generation module is used to generate an interpretable model activation diagram based on the image features;
[0036] The classification module is used to determine the category and corresponding score of the chest X-ray image based on the image features.
[0037] Thirdly, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus, and the processor includes a graphics processor and a central processing unit.
[0038] Memory, used to store computer programs;
[0039] When a processor executes a program stored in memory, it implements the steps of a chest X-ray screening classification prediction method according to any embodiment of the first aspect.
[0040] Fourthly, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of a chest X-ray screening classification prediction method as described in any embodiment of the first aspect.
[0041] The technical solutions provided in the embodiments of the present invention have the following advantages compared with the prior art:
[0042] This invention provides a chest X-ray screening classification prediction method. The method involves enhancing the contrast of the acquired chest X-ray image to obtain a contrast-enhanced image. This enhanced image is then input into a pre-trained classification model to determine the category and corresponding score of the chest X-ray image, as well as a corresponding interpretability model activation diagram. This approach can quickly separate "detected" and "undetected" findings in a chest X-ray, improving doctors' work efficiency and reducing misdiagnosis and missed diagnosis. Attached Figure Description
[0043] Figure 1 This is a schematic diagram of a chest X-ray screening classification and prediction method provided in an embodiment of the present invention;
[0044] Figure 2 This is a schematic diagram of the contrast enhancement process for a chest X-ray image.
[0045] Figure 3 This is a schematic diagram of the screening, classification, and prediction process;
[0046] Figure 4 A schematic diagram of a chest X-ray screening classification prediction device provided in the embodiments of the invention;
[0047] Figure 5 This is a schematic diagram of an electronic device structure provided in an embodiment of the present invention. Detailed Implementation
[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0049] To facilitate understanding of the embodiments of the present invention, further explanations and descriptions will be provided below with reference to the accompanying drawings and specific embodiments. These embodiments do not constitute a limitation on the embodiments of the present invention.
[0050] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element. Additionally, in the description of this invention, "a plurality of" means two or more, unless otherwise expressly specified.
[0051] To address the technical problems mentioned in the background section, embodiments of the present invention provide a chest X-ray screening classification prediction method, executed by a chest X-ray screening classification prediction system, as detailed below. Figure 1 , Figure 1 A chest X-ray screening classification method provided in this embodiment of the invention, such as... Figure 1 As shown, the chest X-ray screening classification method includes the following steps:
[0052] Step 110: Perform contrast enhancement processing on the acquired chest X-ray image to be screened to obtain a contrast-enhanced image.
[0053] The chest X-ray image to be screened is a standard chest X-ray image (CXR). Before contrast enhancement processing, the standard chest X-ray image needs to be converted from 16-bit DICOM format to 8-bit PNG format, with pixel values ranging from 0 to 255. There are significant differences in pixel values among different tissues and organs in the chest X-ray image. To enhance image contrast, contrast enhancement processing is performed to obtain a contrast-enhanced image. Specifically, the image contrast is adaptively enhanced based on the pixel characteristics of the chest X-ray image. For detailed processing procedures, please refer to [link to documentation]. Figure 2 , Figure 2 This is a schematic diagram of the contrast enhancement process for a chest X-ray image, as shown below. Figure 2 As shown, the contrast enhancement process for chest X-ray images includes the following steps:
[0054] Step 210: Estimate the histogram distribution of the chest X-ray image and calculate the cumulative distribution function value.
[0055] Step 220: Estimate the boundary value of contrast stretching of the chest X-ray image based on the cumulative distribution function value, and use the boundary value to perform contrast stretching processing on the chest X-ray image to determine the contrast stretched image.
[0056] Step 230: Determine the radius of the disk-shaped morphological structural unit based on the size of the chest X-ray image.
[0057] Specifically, the radius of the disk-shaped morphological structural unit is determined according to the following formula:
[0058] R = Max(sx, xy) * 3%
[0059] Where sx and sy represent the length and width of the chest X-ray image, respectively, and Max represents taking the maximum value of the length and width.
[0060] Step 240: Based on the radius of the disk-shaped morphological structural unit, perform morphological top-hat filtering and morphological bottom-hat filtering on the contrast-stretched image respectively to determine the morphological top-hat filtered image and the morphological bottom-hat filtered image.
[0061] Step 250: Determine the contrast-enhanced image based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image.
[0062] Specifically, based on the radius of the disk-shaped morphological structural unit obtained in step 230, morphological top-hat filtering (img_tophat) and bottom-hat filtering (img_bottomhat) operations are performed on the contrast-stretched image, respectively. The final contrast-enhanced image is obtained by the following formula:
[0063] img_enhance=img_stretch+img_tophat-img_bothhat
[0064] Here, img_enhance represents a contrast-enhanced image, img_stretch represents a contrast-stretched image, img_tophat represents a morphological top-hat filtered image, and img_botthat represents a morphological bottom-hat filtered image.
[0065] In other words, this application determines a contrast-enhanced image based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image:
[0066] Contrast-enhanced chest X-ray image = contrast-stretched image + morphological top-hat filtered image - morphological bottom-hat filtered image.
[0067] As described above, step 110 involves obtaining the optimal contrast-enhanced image from chest X-ray images from different devices and with different parameters, based on the distribution characteristics of the image pixel values, using adaptive adjustment of image grayscale and morphological opening and closing algorithms, to provide input for the subsequent classification model.
[0068] Step 120: Input the contrast-enhanced image into the pre-trained classification model to determine the category and corresponding score of the chest X-ray image, and the corresponding interpretability model activation diagram.
[0069] Specifically, the classification model mainly consists of an image feature extraction module, an activation map generation module, and a classification module. For details on the screening and classification process, please refer to [link to relevant documentation]. Figure 3 , Figure 3 This is a schematic diagram of the screening and classification process.
[0070] The image feature extraction module adopts a deep neural network model, specifically a convolutional neural network body structure (CNN-Body). Specifically, it is based on a deep residual network (ResNet) as the basic structure. Preferably, the ResNet34 structure is selected in this application. It should be noted that other residual network structures can also be selected according to the requirements, and no limitation is made here.
[0071] The parameters in the image feature extraction module are trained using actual clinical chest X-ray images annotated by senior radiologists. The contrast-enhanced images obtained above are input into the trained image feature extraction module, which automatically extracts the image features of the input chest X-ray image. The classification module then determines the category, and the activation map generation module generates an interpretable classification result diagram. Therefore, the final classification result includes the category of the chest X-ray image, its corresponding score, and the corresponding interpretable model activation diagram. Here, the category of the chest X-ray image refers to whether the image is normal (no abnormalities found) or abnormal (abnormalities found). Therefore, the classification head-layout of the classification module in this application is selected as 2 based on the classification objective. The activation map generation module in this application uses a gradient-based class activation map algorithm, such as... Figure 3 As shown in the figure, CAM Layer represents Class Activation Mapping.
[0072] As described above, the specific functions of the image feature extraction module, activation map generation module, and classification module are as follows:
[0073] The image feature extraction module is used to extract image features from the input contrast-enhanced image.
[0074] The activation graph generation module is used to generate an interpretable model activation diagram based on the image features.
[0075] The classification module is used to determine the category and corresponding score of the chest X-ray image based on the image features.
[0076] Furthermore, the classification model needs to be described. In this application, before application, the classification model utilizes data screening by radiologists and retrospective review by senior physicians to determine the category of chest X-ray images. For example, a training dataset of nearly 20,000 high-quality continuous clinical chest X-ray images (9765 normal and 9956 abnormal) is used to train the pre-constructed classification model. Its feasibility and effectiveness are verified using actual clinical data, ultimately determining the area under the ROC curve for the overall classification performance of the model to be 0.96. That is, the classification model in this application has high accuracy.
[0077] As described above, the chest X-ray screening classification prediction system provided in this application consists of multiple modules and can adapt to actual clinical scenarios. First, it automatically and adaptively processes image characteristics from different image acquisition devices under different imaging conditions, ensuring that the AI model has good generalization ability. Furthermore, the Grad-CAM method is used to obtain an activation heatmap by weighted summation of feature maps generated from convolutional layers of different categories. This activation heatmap can be used to interpret the model's classification results. The CXR classification module uses Grad-CAM to generate a classification activation heatmap, which shows which regions in the input image are important activation regions for obtaining the classification result. Simultaneously, it generates a model activation area map that corresponds one-to-one with the original chest X-ray image of the lesion region, providing interpretable information for clinical use.
[0078] This invention provides a chest X-ray screening classification prediction method. The method involves enhancing the contrast of the acquired chest X-ray image to obtain a contrast-enhanced image. This enhanced image is then input into a pre-trained classification model to determine the category and corresponding score of the chest X-ray image, as well as a corresponding interpretability model activation diagram. This approach can quickly separate "detected" and "undetected" findings in a chest X-ray, improving doctors' work efficiency and reducing misdiagnosis and missed diagnosis.
[0079] The above describes an embodiment of a chest X-ray screening classification prediction method provided by the present invention. Other embodiments provided by the present invention will be described below.
[0080] Figure 4 A schematic diagram of a chest X-ray screening classification prediction device is provided for an embodiment of the invention, as shown below. Figure 4 As shown, the device includes a contrast intensity processing module 401 and a screening classification prediction module 402.
[0081] The contrast intensity processing module 401 is used to perform contrast enhancement processing on the acquired chest X-ray image to be screened to obtain a contrast-enhanced image.
[0082] The screening classification prediction module 402 is used to input the contrast-enhanced image into a pre-trained classification model to determine the category and corresponding score of the chest X-ray image, and the corresponding interpretability model activation diagram.
[0083] In one possible implementation, the contrast intensity processing module 401 is specifically used for:
[0084] Estimate the histogram distribution of the chest X-ray image and calculate the cumulative distribution function value;
[0085] The boundary value of the contrast stretching of the chest X-ray image is estimated based on the cumulative distribution function value, and the contrast stretching of the chest X-ray image is performed using the boundary value to determine the contrast stretched image.
[0086] The radius of the disc-shaped morphological structural unit is determined based on the size of the chest X-ray image;
[0087] Based on the radius of the disk-shaped morphological structural unit, morphological top-hat filtering and morphological bottom-hat filtering are performed on the contrast-stretched image to determine the morphological top-hat filtered image and the morphological bottom-hat filtered image.
[0088] A contrast-enhanced image is determined based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image.
[0089] In one possible implementation, the classification model consists of an image feature extraction module, an activation map generation module, and a classification module, wherein,
[0090] The image feature extraction module is used to extract image features from the input contrast-enhanced image;
[0091] The activation map generation module is used to generate an interpretable model activation diagram based on the image features;
[0092] The classification module is used to determine the category and corresponding score of the chest X-ray image based on the image features.
[0093] The functions performed by each component in the chest X-ray screening classification prediction device provided in this embodiment have been described in detail in any of the above method embodiments, and therefore will not be repeated here.
[0094] This invention provides a chest X-ray screening classification prediction device that performs contrast enhancement processing on the acquired chest X-ray image to be screened, obtaining a contrast-enhanced image. The contrast-enhanced image is then input into a pre-trained classification model to determine the category and corresponding score of the chest X-ray image, as well as a corresponding interpretability model activation diagram. This method can quickly separate "detected" and "undetected" findings in a chest X-ray, improving doctors' work efficiency and reducing misdiagnosis and missed diagnosis.
[0095] like Figure 5 As shown, an embodiment of the present invention provides an electronic device, including a processor 131, a communication interface 132, a memory 133, and a communication bus 134. The processor 131, the communication interface 132, and the memory 133 communicate with each other through the communication bus 134. The processor 131 includes a graphics processor and a central processing unit.
[0096] Memory 133 is used to store computer programs;
[0097] In one embodiment of the present invention, when the processor 131 executes the program stored in the memory 133, it implements the steps of the chest X-ray screening classification prediction method provided in any of the foregoing method embodiments.
[0098] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of a chest X-ray screening classification prediction method as provided in any of the foregoing method embodiments.
[0099] Chest X-ray is the preferred imaging method for respiratory diseases and is the most common clinical task in routine examinations or screenings in medical imaging departments or health check-up centers. In general hospitals, most routine physical examinations and respiratory disease screenings use CXR as the first choice. Therefore, CXR is usually the single examination with the largest workload in outpatient clinics, and a considerable number of imaging diagnostic results in these diagnostic tasks are "no findings". This application proposes a chest X-ray screening classification and prediction method based on deep learning algorithms using CXR, which can quickly separate "found" and "no found" results from CXR in outpatient, emergency, preoperative routine, and physical examination populations, improving the work efficiency of clinicians, reducing missed diagnoses and misdiagnoses, and also providing possibilities for quality control in telemedicine or health check-up centers. It has significant commercial and clinical translation prospects and is worth promoting.
[0100] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0101] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented in hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.
[0102] The above are merely specific embodiments of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.
Claims
1. A chest X-ray screening classification and prediction method, characterized in that, The method includes the following steps: The acquired chest X-ray images to be screened are subjected to contrast enhancement processing to obtain contrast-enhanced images; The contrast-enhanced image is input into a pre-trained classification model to determine the category and corresponding score of the chest X-ray image, as well as the corresponding interpretability model activation diagram; this includes the following steps: Estimate the histogram distribution of the chest X-ray image and calculate the cumulative distribution function value; The boundary value of the contrast stretching of the chest X-ray image is estimated based on the cumulative distribution function value, and the contrast stretching of the chest X-ray image is performed using the boundary value to determine the contrast stretched image. The radius of the disc-shaped morphological structural unit is determined based on the size of the chest X-ray image; Based on the radius of the disk-shaped morphological structural unit, morphological top-hat filtering and morphological bottom-hat filtering are performed on the contrast-stretched image to determine the morphological top-hat filtered image and the morphological bottom-hat filtered image. A contrast-enhanced image is determined based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image.
2. The method according to claim 1, characterized in that, The radius of the disk-shaped morphological structural unit is determined using the following formula: R = Max(sx, xy) × 3%; Where sx and sy represent the length and width of the chest X-ray image, respectively, and Max represents the maximum value of the length and width.
3. The method according to claim 1, wherein a contrast-enhanced image is determined based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image using the following formula: Contrast-enhanced chest X-ray image = contrast-stretched image + morphological top-hat filtered image - morphological bottom-hat filtered image.
4. The method according to claim 1, characterized in that, The classification model consists of an image feature extraction module, an activation map generation module, and a classification module. The image feature extraction module is used to extract image features from the input contrast-enhanced image; The activation map generation module is used to generate an interpretable model activation diagram based on the image features; The classification module is used to determine the category and corresponding score of the chest X-ray image based on the image features.
5. A chest X-ray screening classification prediction device, characterized in that, The device includes: The contrast intensity processing module is used to perform contrast enhancement processing on the acquired chest X-ray images to be screened to obtain contrast-enhanced images; The screening classification prediction module is used to input the contrast-enhanced image into a pre-trained classification model to determine the category and corresponding score of the chest X-ray image, and the corresponding interpretability model activation diagram. The contrast intensity processing module is specifically used for: Estimate the histogram distribution of the chest X-ray image and calculate the cumulative distribution function value; The boundary value of the contrast stretching of the chest X-ray image is estimated based on the cumulative distribution function value, and the contrast stretching of the chest X-ray image is performed using the boundary value to determine the contrast stretched image. The radius of the disc-shaped morphological structural unit is determined based on the size of the chest X-ray image; Based on the radius of the disk-shaped morphological structural unit, morphological top-hat filtering and morphological bottom-hat filtering are performed on the contrast-stretched image to determine the morphological top-hat filtered image and the morphological bottom-hat filtered image. A contrast-enhanced image is determined based on the contrast-stretched image, the morphological top-hat filtered image, and the morphological bottom-hat filtered image.
6. The apparatus according to claim 5, characterized in that, The classification model consists of an image feature extraction module, an activation map generation module, and a classification module. The image feature extraction module is used to extract image features from the input contrast-enhanced image; The activation map generation module is used to generate an interpretable model activation diagram based on the image features; The classification module is used to determine the category and corresponding score of the chest X-ray image based on the image features.
7. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other through the communication bus. The processor includes a graphics processor and a central processing unit. Memory, used to store computer programs; The processor, when executing a program stored in memory, implements the steps of a chest X-ray screening classification prediction method as described in any one of claims 1 to 4.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of a chest X-ray screening classification prediction method as described in any one of claims 1 to 4.