Virus detection method applied to X-ray image and related equipment

By using AtomNAS algorithm and multi-scale denoising algorithm in virus detection, the problem of insufficient accuracy in capturing pneumonia X-ray images has been solved, and higher detection accuracy and faster detection process have been achieved , reducing medical costs.

CN119067945APending Publication Date: 2024-12-03PING AN TECH (SHENZHEN) CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411170534.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-23
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Traditional virus detection methods have low accuracy when capturing key features in pneumonia X-ray images, resulting in the impact of the accuracy of virus detection and recognition.

Method used

The AtomNAS algorithm is used to generate the initial neural architecture search model in the neural network architecture search space, and the X-ray image is preprocessed through multi-scale denoising algorithm and adaptive histogram equalization technology. Combined with random occlusion and parameter optimization, the training data is constructed to Improve model performance.

Benefits of technology

It significantly improves the accuracy of virus testing, reduces the work burden of doctors, reduces the possibility of human diagnosis errors, and provides patients with faster and more accurate testing services, reducing medical costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119067945A_ABST
    Figure CN119067945A_ABST
Patent Text Reader

Abstract

The embodiment of the invention belongs to the technical field of image processing, and relates to a virus detection method applied to an X-ray image and related equipment.The method comprises the steps that a search space containing multiple neural network architectures is constructed by collecting historical lung X-ray photos and virus detection results corresponding to the historical lung X-ray photos, an initial neural architecture search model is generated in the search space using a neural architecture search (NAS) algorithm, exclusively for identifying and predicting viral infections in X-ray images. According to the method, the workload of doctors can be greatly reduced, the possibility of artificial diagnosis errors is reduced, and a faster and more accurate virus detection service is provided for patients; in addition, the medical cost is reduced, and the overall operation efficiency of a hospital system is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a virus detection method and related equipment applied to X-ray images. Background Art

[0002] In recent years, human medical level has faced new challenges. Pneumonia is mainly manifested by fever, dry cough, fatigue, etc. A small number of patients are accompanied by upper respiratory and digestive tract symptoms such as nasal congestion, runny nose, diarrhea, etc. At present, the main transmission routes of pneumonia that can be determined are direct transmission, aerosol transmission and contact transmission. Direct transmission refers to the infection caused by the patient's sneezing, coughing, talking droplets, and exhaled gas directly inhaled at close range; aerosol transmission refers to the droplets mixing in the air to form aerosols, which cause infection after inhalation; contact transmission refers to the droplets deposited on the surface of objects, contacting contaminated hands, and then contacting the oral, nasal, eye and other mucous membranes, causing infection. In the process of epidemic prevention, early detection and isolation of the epidemic play a vital role.

[0003] There is an existing virus detection method that uses a convolutional neural network to capture key features in X-ray images to achieve virus detection in patients.

[0004] However, the applicant found that because the characteristics of pneumonia X-ray images are relatively hidden and easily confused with other lung diseases, traditional virus detection methods are difficult to accurately capture key features, which in turn affects the accuracy of virus detection / identification. Summary of the invention

[0005] The purpose of the embodiments of the present application is to provide a virus detection method and related equipment applied to X-ray images to solve the problem of low accuracy of traditional virus detection methods.

[0006] In order to solve the above technical problems, the present application embodiment provides a virus detection method applied to X-ray images, which adopts the following technical solutions:

[0007] Reading a database, obtaining historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database;

[0008] Constructing a search space including several neural network architectures, and generating an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm;

[0009] Constructing model training data and model verification data based on the historical lung X-ray images and the historical virus detection results;

[0010] Inputting the model training data into the initial neural architecture search model to perform a model training operation, and optimizing parameters of the model training operation according to the model verification data to obtain a trained neural architecture search model;

[0011] Receiving a virus detection request carrying a picture to be detected sent by a user terminal;

[0012] Performing a virus recognition operation on the image to be detected according to the trained neural architecture search model to obtain a virus recognition result;

[0013] The virus identification result is output to the user terminal.

[0014] Furthermore, after the step of reading the database and acquiring the historical lung X-ray images and the historical virus detection results corresponding to the historical lung X-ray images in the database, the following steps are also included:

[0015] Calculating wavelet coefficients of the historical lung X-ray image according to a multi-scale denoising algorithm;

[0016] Classifying the wavelet coefficients according to the frequency characteristics of the historical lung X-ray images to obtain classified wavelet coefficients;

[0017] A soft threshold shrinkage operation is performed on the classified wavelet coefficients according to an adaptive threshold method to obtain a denoised historical lung X-ray image.

[0018] Furthermore, after the step of performing a soft threshold shrinkage operation on the classified wavelet coefficients according to the adaptive threshold method to obtain a denoised historical lung X-ray image, the following steps are also included:

[0019] Performing quality assessment on the contrast and clarity of the denoised historical lung X-ray image to obtain a quality assessment result;

[0020] If the quality assessment result fails, adjusting the parameters of the multi-scale denoising algorithm;

[0021] The wavelet coefficients of the historical lung X-ray image are recalculated according to the multi-scale denoising algorithm with adjusted parameters, and the historical lung X-ray image is denoised according to the new wavelet parameters until the re-denoised historical lung X-ray image meets the preset contrast and clarity.

[0022] Furthermore, after the step of reading the database and obtaining the historical lung X-ray images and the historical virus detection results corresponding to the historical lung X-ray images in the database, the following steps are also included:

[0023] performing a contrast enhancement operation on the historical lung X-ray image according to an adaptive histogram equalization technique to obtain an enhanced historical lung X-ray image;

[0024] Generate a rectangular occlusion area according to a random algorithm, and perform a random occlusion operation on the enhanced historical lung X-ray image according to the rectangular occlusion area to obtain the historical lung X-ray image after random occlusion processing;

[0025] Analyzing feature distribution data of the historical lung X-ray images after the random occlusion processing;

[0026] The parameter setting of the adaptive histogram equalization technology is adjusted according to the characteristic distribution data, and the image enhancement operation is performed on the historical lung X-ray image according to the adaptive histogram equalization technology after the parameter adjustment.

[0027] Furthermore, the step of constructing a search space including several neural network architectures and generating an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm specifically includes the following steps:

[0028] Acquire, in the search space according to the AtomNAS algorithm, a plurality of high-performance model structures corresponding to the historical virus detection results;

[0029] The high-performance model structures are stacked according to the block method to construct the initial neural architecture search model, wherein each of the high-performance model structures is bound to a scaling amount α, so that the initial neural architecture search model updates the scaling amount α after each round of training.

[0030] In order to solve the above technical problems, the embodiment of the present application further provides a virus detection device applied to X-ray images, which adopts the following technical solution:

[0031] A historical image acquisition module, used to read a database, and acquire historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database;

[0032] A model building module, used to build a search space containing several neural network architectures, and generate an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm;

[0033] A training data construction module, used to construct model training data and model verification data based on the historical lung X-ray images and the historical virus detection results;

[0034] A model training module, used to input the model training data into the initial neural architecture search model to perform a model training operation, and optimize the parameters of the model training operation according to the model verification data to obtain a trained neural architecture search model;

[0035] A request acquisition module, used to receive a virus detection request carrying a picture to be detected sent by a user terminal;

[0036] An identification acquisition module, used to perform a virus identification operation on the image to be detected according to the trained neural architecture search model to obtain a virus identification result;

[0037] The result output module is used to output the virus identification result to the user terminal.

[0038] Furthermore, the device also includes:

[0039] A wavelet coefficient calculation module, used for calculating the wavelet coefficients of the historical lung X-ray image according to a multi-scale denoising algorithm;

[0040] A wavelet coefficient classification module, used for classifying the wavelet coefficients according to the frequency characteristics of the historical lung X-ray images to obtain classified wavelet coefficients;

[0041] The soft threshold shrinkage module is used to perform a soft threshold shrinkage operation on the classified wavelet coefficients according to an adaptive threshold method to obtain a denoised historical lung X-ray image.

[0042] Furthermore, the device also includes:

[0043] A quality assessment module, used to perform quality assessment on the contrast and clarity of the denoised historical lung X-ray image to obtain a quality assessment result;

[0044] A parameter adjustment module, used for adjusting the parameters of the multi-scale denoising algorithm if the quality assessment result fails;

[0045] The denoising module is used to recalculate the wavelet coefficients of the historical lung X-ray image according to the multi-scale denoising algorithm with adjusted parameters, and perform denoising operation on the historical lung X-ray image according to the new wavelet parameters until the obtained re-denoised historical lung X-ray image meets the preset contrast and clarity.

[0046] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0047] The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the virus detection method applied to X-ray images as described above when executing the computer-readable instructions.

[0048] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0049] The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps of the virus detection method applied to X-ray images as described above are implemented.

[0050] The present application provides a virus detection method applied to X-ray images, including: reading a database, obtaining historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database; constructing a search space containing several neural network architectures, and generating an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm; constructing model training data and model verification data according to the historical lung X-ray images and the historical virus detection results; inputting the model training data into the initial neural architecture search model for model training operation, and optimizing the parameters of the model training operation according to the model verification data to obtain a trained neural architecture search model; receiving a virus detection request carrying a picture to be detected sent by a user terminal; performing a virus recognition operation on the picture to be detected according to the trained neural architecture search model to obtain a virus recognition result; and outputting the virus recognition result to the user terminal. Compared with the prior art, the present application can greatly reduce the workload of doctors, reduce the possibility of human diagnostic errors, and provide patients with faster and more accurate virus detection services; in addition, it also helps to reduce medical costs and improve the overall operational efficiency of hospital systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0053] Figure 2 is a flowchart of the implementation of the virus detection method applied to X-ray images provided in the first embodiment of the present application;

[0054] Figure 3 is a schematic structural diagram of a virus detection device applied to X-ray images provided in Example 2 of the present application;

[0055] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0057] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0058] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0059] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0060] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0061] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.

[0062] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0063] It should be noted that the virus detection method applied to X-ray images provided in the embodiments of the present application is generally executed by a server / terminal device, and accordingly, the virus detection device applied to X-ray images is generally set in the server / terminal device.

[0064] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0065] Embodiment 1

[0066] Continue to refer Figure 2 , shows a flow chart of an embodiment of a virus detection method applied to X-ray images according to the present application. The virus detection method applied to X-ray images comprises: step S201, step S202, step S203, step S204, step S205, step S206 and step S207.

[0067] In step S201, a database is read to obtain historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database.

[0068] In an embodiment of the present application, relevant data of historical virus detection is stored in the database, specifically including historically collected lung X-rays and information on whether the patient is infected with the virus.

[0069] In step S202, a search space including several neural network architectures is constructed, and an initial neural architecture search model corresponding to historical virus detection results is generated in the search space according to the AtomNAS algorithm.

[0070] In the embodiment of the present application, a search space needs to be pre-built, which contains a variety of possible neural network architectures. Then, the NAS algorithm is used to generate candidate network architectures in this search space, and the performance of these architectures is evaluated through training and verification. Finally, through multiple iterations and selections, a network architecture that performs high performance in the problem of virus recognition and classification of lung X-rays is searched.

[0071] In step S203, model training data and model verification data are constructed based on historical lung X-ray images and historical virus detection results.

[0072] In an embodiment of the present application, several data sets are formed according to the correlation between historical lung X-ray photographs and historical virus detection results, and the data sets are divided into training data and verification data by a preset division method.

[0073] In step S204, the model training data is input into the initial neural architecture search model for model training operation, and the parameters of the model training operation are optimized according to the model verification data to obtain a trained neural architecture search model.

[0074] In the embodiment of the present application, the optimal neural network architecture is used in the training process, and supervised learning is performed using a labeled data set. During the training process, the model learns how to extract features from the input image and map these features to corresponding category labels.

[0075] In the embodiment of the present application, the performance of the trained model is evaluated on an independent test set. The evaluation indicators usually include accuracy, recall, F1 score, etc. If the model performance meets the requirements, it can be applied to the actual scenario.

[0076] In step S205, a virus detection request carrying a picture to be detected is received from a user terminal.

[0077] In the embodiments of the present application, the user terminal refers to a terminal device used to execute the image processing method for preventing document abuse provided by the present application. The user terminal can be a mobile terminal such as a mobile phone, a smart phone, a laptop computer, a digital broadcast receiver, a PDA (personal digital assistant), a PAD (tablet computer), a PMP (portable multimedia player), a navigation device, etc., as well as a fixed terminal such as a digital TV, a desktop computer, etc. It should be understood that the examples of user terminals here are only for the convenience of understanding and are not used to limit the present application.

[0078] In step S206, a virus recognition operation is performed on the image to be detected according to the trained neural architecture search model to obtain a virus recognition result.

[0079] In step S207, the virus identification result is output to the user terminal.

[0080] In an embodiment of the present application, a virus detection method applied to X-ray images is provided, including: reading a database, obtaining historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database; constructing a search space containing several neural network architectures, and generating an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm; constructing model training data and model verification data according to the historical lung X-ray images and the historical virus detection results; inputting the model training data into the initial neural architecture search model for model training operation, and optimizing the parameters of the model training operation according to the model verification data to obtain a trained neural architecture search model; receiving a virus detection request carrying a picture to be detected sent by a user terminal; performing a virus recognition operation on the picture to be detected according to the trained neural architecture search model to obtain a virus recognition result; and outputting the virus recognition result to the user terminal. Compared with the prior art, the present application can greatly reduce the workload of doctors, reduce the possibility of human diagnostic errors, and provide patients with faster and more accurate virus detection services; in addition, it also helps to reduce medical costs and improve the overall operational efficiency of the hospital system.

[0081] In some optional implementations of the embodiments of the present application, after step S201, the following steps are further included:

[0082] Calculate the wavelet coefficients of historical lung X-ray images based on a multi-scale denoising algorithm;

[0083] Classify the wavelet coefficients according to the frequency characteristics of the historical lung X-ray images to obtain the classified wavelet coefficients;

[0084] The classified wavelet coefficients are subjected to soft threshold shrinkage operation according to the adaptive threshold method to obtain the denoised historical lung X-ray image.

[0085] In an embodiment of the present application, when a multi-scale denoising algorithm based on wavelet transform is used to process an X-ray image, multi-scale information of the image is first extracted by wavelet transform. Specifically, a three-layer decomposition using Daubechies wavelet basis can be selected to obtain details and approximate coefficients.

[0086] In the embodiment of the present application, these wavelet coefficients are classified according to the frequency characteristics they represent, and high-frequency and low-frequency coefficients are processed separately, wherein the high-frequency coefficients usually contain noise and edge information, while the low-frequency coefficients contain more image main body information.

[0087] In the embodiment of the present application, an adaptive threshold method is applied to the wavelet coefficients of each scale for soft threshold processing, such as setting the threshold to half of the standard deviation of each coefficient, so as to effectively remove noise and retain important details.

[0088] In an embodiment of the present application, by denoising historical lung X-ray images, the influence of invalid features on subsequent model training is effectively suppressed, and the recognition accuracy of the neural architecture search model is effectively improved.

[0089] In some optional implementations of the embodiments of the present application, after the step of performing a soft threshold shrinkage operation on the classified wavelet coefficients according to the adaptive threshold method to obtain a denoised historical lung X-ray image, the following steps are also included:

[0090] Performing quality assessment on the contrast and clarity of the denoised historical lung X-ray images to obtain a quality assessment result;

[0091] If the quality assessment result fails, adjust the parameters of the multi-scale denoising algorithm;

[0092] The wavelet coefficients of the historical lung X-ray images are recalculated according to the multi-scale denoising algorithm with adjusted parameters, and the historical lung X-ray images are denoised according to the new wavelet parameters until the re-denoised historical lung X-ray images meet the preset contrast and clarity.

[0093] In an embodiment of the present application, after denoising, the image can be reconstructed using an inverse wavelet transform, and then the quality of the reconstructed image is evaluated, including calculation of contrast and clarity, for example, using peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) as evaluation criteria.

[0094] In an embodiment of the present application, if the image after preliminary denoising does not achieve the expected effect, the parameters of the wavelet transform will be adjusted, such as selecting a different wavelet basis or adjusting the number of decomposition layers, and then the denoising process will be performed again until the re-denoised image achieves the expected contrast and clarity.

[0095] In the embodiments of the present application, the final denoised image will be evaluated based on the overall visual effect to ensure that the image quality meets the requirements of subsequent analysis and processing. At the same time, the entire denoising process is completely based on automatic processing based on information technology, which effectively reduces human intervention and improves processing efficiency and consistency of image quality.

[0096] In some optional implementations of the embodiments of the present application, after obtaining a re-denoised image that achieves the expected contrast and clarity, a Canny edge detection algorithm can be used to confirm whether the edges of the image are clearer.

[0097] In some optional implementations of the embodiments of the present application, after obtaining a re-denoised image that achieves the expected contrast and clarity, image texture analysis techniques such as gray level co-occurrence matrix (GLCM) can also be used to analyze the clarity and uniformity of the texture to ensure that important structural information in the image is not lost.

[0098] In the embodiment of the present application, multi-level quality assessment is performed to ensure that each step of processing is accurate and effective, thereby achieving the optimal image processing effect.

[0099] In some optional implementations of the embodiments of the present application, after the step of performing a soft threshold shrinkage operation on the classified wavelet coefficients according to the adaptive threshold method to obtain a denoised historical lung X-ray image, the following steps are also included:

[0100] Performing contrast enhancement operation on the historical lung X-ray image according to the adaptive histogram equalization technology to obtain an enhanced historical lung X-ray image;

[0101] Generate a rectangular occlusion region according to a random algorithm, and perform a random occlusion operation on the enhanced historical lung X-ray image according to the rectangular occlusion region to obtain the historical lung X-ray image after random occlusion processing;

[0102] Analyze the feature distribution data of historical lung X-ray images after random occlusion processing;

[0103] The parameter setting of the adaptive histogram equalization technology is adjusted according to the characteristic distribution data, and the image enhancement operation is performed on the historical lung X-ray image according to the adaptive histogram equalization technology after the parameter adjustment.

[0104] In an embodiment of the present application, in the process of processing X-ray images, adaptive histogram equalization technology is first used to enhance the contrast of the lung area in the image. Specifically, the histogram of each small area is calculated, and then the brightness of the image is adjusted according to these histograms to improve the local visibility and detail resolution of the image.

[0105] In an embodiment of the present application, after contrast enhancement, a random algorithm is used to generate multiple rectangular occlusion areas in the image with random sizes and positions to simulate possible image occlusion problems in actual situations, such as randomly selecting 5% of the area in the image for occlusion.

[0106] In an embodiment of the present application, a bicubic interpolation algorithm is applied to perform elastic deformation operations on the image to simulate the deformation caused by changes in shooting angle and body position. This operation can be performed by randomly generating deformation parameters (such as deformation coefficients of 1 to 3 in the horizontal and vertical directions) to increase the diversity of image processing.

[0107] In an embodiment of the present application, the image processed by these enhancement operations is analyzed using image processing software to analyze the feature distribution of the enhanced image, such as histogram analysis and texture analysis, and the parameter settings in the above-mentioned data enhancement strategy are adjusted based on this, such as adjusting the block size of the histogram equalization or the range of the deformation coefficient, to ensure the adaptability and optimization of the strategy.

[0108] In an embodiment of the present application, after determining the optimal parameter settings, these parameters are included in the search space, and the strategy combination is optimized using automatic search techniques such as genetic algorithms to find the parameter combination that can maximize the image quality and model performance.

[0109] In an embodiment of the present application, the searched optimal parameter combination is used to perform local contrast enhancement, random occlusion, elastic deformation and other operations on the X-ray image again to generate a final enhanced image data set.

[0110] In an embodiment of the present application, using this batch of enhanced image data sets to train a deep learning model can effectively improve the model's recognition ability and accuracy of X-ray image features.

[0111] In some optional implementations of the embodiments of the present application, the above step S202 specifically includes the following steps:

[0112] According to the AtomNAS algorithm, several high-performance model structures corresponding to historical virus detection results are obtained in the search space;

[0113] The high-performance model structures are stacked according to the block method to construct the initial neural architecture search model, wherein each high-performance model structure is bound to a scaling amount α so that the initial neural architecture search model updates the scaling amount α after each round of training.

[0114] In the embodiment of the present application, the NAS model used is AtomNAS, which is based on the block method. It uses two connected convolutional layers as the basic block and continuously stacks blocks to form the final network structure. In the network structure, each block is bundled with a scaling amount, which will be updated after each round of training.

[0115] In some optional implementations of the embodiments of the present application, in order to make the final model not too large, the model also uses a model pruning method. The scaling parameter will converge with each round of training. When it is too small, the bound block will be considered invalid and pruned, thereby reducing the size of the model, making the model closer to the actual application scenario and improving the detection speed.

[0116] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0117] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0118] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0119] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0120] Embodiment 2

[0121] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a virus detection device applied to X-ray images. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0122] like Figure 3As shown, the virus detection device 200 applied to X-ray images of the embodiment of the present application includes:

[0123] A historical image acquisition module 210 is used to read a database and acquire historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database;

[0124] A model building module 220, for building a search space including several neural network architectures, and generating an initial neural architecture search model corresponding to historical virus detection results in the search space according to the AtomNAS algorithm;

[0125] A training data construction module 230, for constructing model training data and model verification data based on historical lung X-ray images and historical virus detection results;

[0126] A model training module 240 is used to input the model training data into the initial neural architecture search model to perform a model training operation, and optimize the parameters of the model training operation according to the model verification data to obtain a trained neural architecture search model;

[0127] The request acquisition module 250 is used to receive a virus detection request carrying a picture to be detected sent by a user terminal;

[0128] The recognition acquisition module 260 is used to perform a virus recognition operation on the image to be detected according to the trained neural architecture search model to obtain a virus recognition result;

[0129] The result output module 270 is used to output the virus identification result to the user terminal.

[0130] In an embodiment of the present application, a virus detection device 200 applied to X-ray images is provided, including: a historical image acquisition module 210, which is used to read a database and obtain historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database; a model construction module 220, which is used to construct a search space containing several neural network architectures, and generate an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm; a training data construction module 230, which is used to construct model training data and model verification data according to historical lung X-ray images and historical virus detection results; a model training module 240, which is used to input the model training data into the initial neural architecture search model for model training operation, and optimize the parameters of the model training operation according to the model verification data to obtain a trained neural architecture search model; a request acquisition module 250, which is used to receive a virus detection request carrying a picture to be detected sent by a user terminal; an identification acquisition module 260, which is used to perform a virus identification operation on the picture to be detected according to the trained neural architecture search model to obtain a virus identification result; and a result output module 270, which is used to output the virus identification result to the user terminal. Compared with the existing technology, this application can greatly reduce the workload of doctors, reduce the possibility of human diagnostic errors, and provide patients with faster and more accurate virus testing services; in addition, it can also help reduce medical costs and improve the overall operational efficiency of the hospital system.

[0131] In some optional implementations of the embodiments of the present application, the above-mentioned virus detection device 200 applied to X-ray images further includes:

[0132] A wavelet coefficient calculation module, used to calculate the wavelet coefficients of the historical lung X-ray images according to the multi-scale denoising algorithm;

[0133] A wavelet coefficient classification module is used to classify the wavelet coefficients according to the frequency characteristics of the historical lung X-ray images to obtain the classified wavelet coefficients;

[0134] The soft threshold shrinkage module is used to perform a soft threshold shrinkage operation on the classified wavelet coefficients according to an adaptive threshold method to obtain a denoised historical lung X-ray image.

[0135] In some optional implementations of the embodiments of the present application, the above-mentioned virus detection device 200 applied to X-ray images further includes:

[0136] A quality assessment module is used to perform quality assessment on the contrast and clarity of the denoised historical lung X-ray images to obtain a quality assessment result;

[0137] A parameter adjustment module, used to adjust the parameters of the multi-scale denoising algorithm if the quality assessment result fails;

[0138] The denoising module is used to recalculate the wavelet coefficients of the historical lung X-ray image according to the multi-scale denoising algorithm with adjusted parameters, and perform denoising operation on the historical lung X-ray image according to the new wavelet parameters until the obtained re-denoised historical lung X-ray image meets the preset contrast and clarity.

[0139] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of a computer device according to an embodiment of the present application.

[0140] The computer device 300 includes a memory 310, a processor 320, and a network interface 330 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 300 having components 310-330, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0141] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0142] The memory 310 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 310 can be an internal storage unit of the computer device 300, such as a hard disk or memory of the computer device 300. In other embodiments, the memory 310 can also be an external storage device of the computer device 300, such as a plug-in hard disk equipped on the computer device 300, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card, etc. Of course, the memory 310 can also include both the internal storage unit of the computer device 300 and its external storage device. In the embodiment of the present application, the memory 310 is generally used to store the operating system and various application software installed on the computer device 300, such as computer-readable instructions for a virus detection method applied to X-ray images, etc. In addition, the memory 310 can also be used to temporarily store various data that have been output or are to be output.

[0143] The processor 320 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 320 is generally used to control the overall operation of the computer device 300. In the embodiment of the present application, the processor 320 is used to run the computer-readable instructions or process data stored in the memory 310, such as running the computer-readable instructions of the virus detection method applied to X-ray images.

[0144] The network interface 330 may include a wireless network interface or a wired network interface. The network interface 330 is generally used to establish a communication connection between the computer device 300 and other electronic devices.

[0145] The computer equipment provided by this application can greatly reduce the workload of doctors, reduce the possibility of human diagnostic errors, and provide patients with faster and more accurate virus detection services; in addition, it can also help reduce medical costs and improve the overall operational efficiency of the hospital system.

[0146] The present application also provides another embodiment, namely, providing a computer-readable storage medium, which stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the virus detection method applied to X-ray images as described above.

[0147] The computer-readable storage medium provided in this application can greatly reduce the workload of doctors, reduce the possibility of human diagnostic errors, and provide patients with faster and more accurate virus detection services; in addition, it can also help reduce medical costs and improve the overall operational efficiency of the hospital system.

[0148] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0149] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A virus detection method applied to X-ray images, It is characterized in that The steps include: Reading a database, obtaining historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database; Constructing a search space including several neural network architectures, and generating an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm; Constructing model training data and model verification data based on the historical lung X-ray images and the historical virus detection results; Inputting the model training data into the initial neural architecture search model to perform a model training operation, and optimizing parameters of the model training operation according to the model verification data to obtain a trained neural architecture search model; Receiving a virus detection request carrying a picture to be detected sent by a user terminal; Performing a virus recognition operation on the image to be detected according to the trained neural architecture search model to obtain a virus recognition result; The virus identification result is output to the user terminal.

2. The virus detection method applied to X-ray images according to claim 1, It is characterized in that After the step of reading the database and acquiring the historical lung X-ray images and the historical virus detection results corresponding to the historical lung X-ray images in the database, the following steps are also included: Calculating wavelet coefficients of the historical lung X-ray image according to a multi-scale denoising algorithm; Classifying the wavelet coefficients according to the frequency characteristics of the historical lung X-ray images to obtain classified wavelet coefficients; A soft threshold shrinkage operation is performed on the classified wavelet coefficients according to an adaptive threshold method to obtain a denoised historical lung X-ray image.

3. The virus detection method applied to X-ray images according to claim 2, It is characterized in that After the step of performing a soft threshold shrinkage operation on the classified wavelet coefficients according to the adaptive threshold method to obtain a denoised historical lung X-ray image, the following steps are also included: Performing a quality assessment on the contrast and clarity of the denoised historical lung X-ray image to obtain a quality assessment result; If the quality assessment result fails, adjusting the parameters of the multi-scale denoising algorithm; The wavelet coefficients of the historical lung X-ray image are recalculated according to the multi-scale denoising algorithm with adjusted parameters, and the historical lung X-ray image is denoised according to the new wavelet parameters until the re-denoised historical lung X-ray image meets the preset contrast and clarity.

4. The virus detection method applied to X-ray images according to claim 1, It is characterized in that After the step of reading the database and acquiring the historical lung X-ray images and the historical virus detection results corresponding to the historical lung X-ray images in the database, the following steps are also included: performing a contrast enhancement operation on the historical lung X-ray image according to an adaptive histogram equalization technique to obtain an enhanced historical lung X-ray image; Generate a rectangular occlusion area according to a random algorithm, and perform a random occlusion operation on the enhanced historical lung X-ray image according to the rectangular occlusion area to obtain the historical lung X-ray image after random occlusion processing; Analyzing feature distribution data of the historical lung X-ray images after the random occlusion processing; The parameter setting of the adaptive histogram equalization technology is adjusted according to the characteristic distribution data, and the image enhancement operation is performed on the historical lung X-ray image according to the adaptive histogram equalization technology after the parameter adjustment.

5. The virus detection method applied to X-ray images according to claim 1, It is characterized in that The step of constructing a search space including several neural network architectures and generating an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm specifically includes the following steps: Acquire, in the search space according to the AtomNAS algorithm, a plurality of high-performance model structures corresponding to the historical virus detection results; The high-performance model structures are stacked according to the block method to construct the initial neural architecture search model, wherein each of the high-performance model structures is bound to a scaling amount α, so that the initial neural architecture search model updates the scaling amount α after each round of training.

6. A virus detection device applied to X-ray images, It is characterized in that include: A historical image acquisition module, used to read a database, and acquire historical lung X-ray images and historical virus detection results corresponding to the historical lung X-ray images in the database; A model building module, used to build a search space containing several neural network architectures, and generate an initial neural architecture search model corresponding to the historical virus detection results in the search space according to the AtomNAS algorithm; A training data construction module, used to construct model training data and model verification data based on the historical lung X-ray images and the historical virus detection results; A model training module, used to input the model training data into the initial neural architecture search model to perform a model training operation, and optimize the parameters of the model training operation according to the model verification data to obtain a trained neural architecture search model; A request acquisition module, used to receive a virus detection request carrying a picture to be detected sent by a user terminal; An identification acquisition module, used to perform a virus identification operation on the image to be detected according to the trained neural architecture search model to obtain a virus identification result; The result output module is used to output the virus identification result to the user terminal.

7. The virus detection device for X-ray images according to claim 6, It is characterized in that The device also includes: A wavelet coefficient calculation module, used for calculating the wavelet coefficients of the historical lung X-ray image according to a multi-scale denoising algorithm; A wavelet coefficient classification module, used for classifying the wavelet coefficients according to the frequency characteristics of the historical lung X-ray images to obtain classified wavelet coefficients; The soft threshold shrinkage module is used to perform a soft threshold shrinkage operation on the classified wavelet coefficients according to an adaptive threshold method to obtain a denoised historical lung X-ray image.

8. The virus detection device for X-ray images according to claim 7, It is characterized in that The device also includes: A quality assessment module, used to perform quality assessment on the contrast and clarity of the denoised historical lung X-ray image to obtain a quality assessment result; A parameter adjustment module, used for adjusting the parameters of the multi-scale denoising algorithm if the quality assessment result fails; The denoising module is used to recalculate the wavelet coefficients of the historical lung X-ray image according to the multi-scale denoising algorithm with adjusted parameters, and perform denoising operation on the historical lung X-ray image according to the new wavelet parameters until the obtained re-denoised historical lung X-ray image meets the preset contrast and clarity.

9. A computer device comprising a memory and a processor, It is characterized in that The memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the virus detection method applied to X-ray images as described in any one of claims 1 to 5 are implemented.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the virus detection method applied to X-ray images as described in any one of claims 1 to 5.

Citation Information

Cited By

  • A portable pneumonia symptom identification auxiliary analysis system integration device

    TWI934680B