Cervical lesion related examination result analysis method and apparatus, device, and medium

Abstract

Embodiments of the present application relate to the technical field of examination result analysis, and provide a cervical lesion related examination result analysis method and apparatus, a device, and a medium. The method comprises: receiving a case report picture uploaded by a current user terminal, and reading case data in the case report picture; acquiring prediction result information corresponding to the case data on the basis of a current prediction model, and selecting optimal expert information on the basis of the prediction result information; pushing the case data and the prediction result information to an expert terminal corresponding to the optimal expert information, and receiving a final clinical analysis result fed back by the expert terminal; and sending the final clinical analysis result to the current user terminal corresponding to a current user. The present application sends a final clinical analysis result to a current user terminal, so that a current user can conveniently check the final clinical analysis result, and choose whether to perform further pathological examination on the basis of the final clinical analysis result, so that the rationality of medical resource use is improved.

Description

A cervical lesion related examination result analysis method, device, equipment and medium TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of examination result analysis, and more particularly to a cervical lesion related examination result analysis method, device, equipment and medium. BACKGROUND

[0002] According to statistics, there were about 3.929 million new cases of malignant tumors in China in 2015, and about 2.338 million deaths. The overall cancer burden in China is still on the rise, and the cancer prevention and control situation is still severe. Among them, there are about 111,000 new cases of cervical cancer, ranking 6th in female malignant tumors in China. Cervical cancer has become one of the major diseases threatening women's health. At present, the reality of capacity, time, resources and other conditions makes it difficult to quickly improve the standardization level of clinical diagnosis and treatment and the level of medical services, and it is difficult to meet the needs of the people for medical services.

[0003] The existing medical analysis process is: the user performs pathological examination; and then analyzes the cervical lesion situation through the corresponding result after pathological examination. However, among the users in China who receive cervical cancer screening, more than 50% of high-risk users are confirmed by pathological examination to have no high-grade cervical lesions or cancer risk, which means that most users actually do not need to perform pathological examination, thus causing serious waste of medical resources, excessive referral of colposcopy and excessive medical treatment of users.

[0004] SUMMARY

[0005] In order to reduce the waste of medical resources, the embodiments of the present application provide a cervical lesion related examination result analysis method, device, equipment and medium.

[0006] In a first aspect of the present application, a cervical lesion related examination result analysis method is provided.

[0007] A cervical lesion related examination result analysis method comprises:

[0008] receiving a case report picture uploaded by a current user terminal, reading case data in the case report picture;

[0009] obtaining prediction result information corresponding to the case data based on a current prediction model, and selecting optimal expert information based on the prediction result information;

[0010] pushing the case data and the prediction result information to an expert terminal corresponding to the optimal expert information, and receiving a final clinical analysis result fed back by the expert terminal;

[0011] sending the final clinical analysis result to a current user terminal corresponding to the current user.

[0012] By the technical solution, the final clinical analysis result is sent to the current user terminal, the current user can view the final clinical analysis result, and whether to perform further pathological examination is selected according to the final clinical analysis result, thereby improving the rationality of use of medical resources.

[0013] Optionally, before the step of obtaining the prediction result information corresponding to the case data based on the current prediction model, the method further comprises:

[0014] obtaining a current report type corresponding to the case report picture;

[0015] if the current report type is a preset report type related to cervical lesion, obtaining a total number of times of uploading corresponding to the case report picture;

[0016] obtaining a threshold number of times corresponding to the current user, and determining whether the total number of times of uploading reaches the threshold number of times;

[0017] if yes, sending first prompt information to the current user terminal;

[0018] if no, performing the step of obtaining the prediction result information corresponding to the case data based on the current prediction model.

[0019] By the technical solution, the current user is inquired through the first prompt information to determine whether the case report picture is uploaded by mistake, thereby reducing the influence of uploading the case report picture by mistake and improving the accuracy of the prediction result information.

[0020] Optionally, after the step of obtaining the current report type corresponding to the case report picture, the method further comprises:

[0021] based on the current report type, obtaining a current model type corresponding thereto;

[0022] taking a prediction model of the current model type as the current prediction model.

[0023] By the technical solution, when the current user uploads one or more different case report pictures, the electronic device selects a corresponding model type according to one or more report types, and predicts the case data through a prediction model corresponding to the model type, thereby improving the adaptability to different report type combinations possessed by the user.

[0024] Optionally, before the step of reading the case data in the case report picture, the method further comprises:

[0025] if the current report type is not a preset optimal type, obtaining feature information in the case report picture;

[0026] adjust the initial enhancement parameter based on the feature information to obtain a current enhancement parameter;

[0027] perform enhancement processing on the case report picture based on the current enhancement parameter.

[0028] Through the above technical solution, various changes that may occur in the actual shooting process of the case report picture can be simulated by performing enhancement processing on the case report picture, so as to achieve the purpose of data expansion. When the current report type is not the best preset type, data expansion is performed to improve the accuracy of subsequent case data reading and reduce the influence of the non-ideal report type on case data reading.

[0029] Optionally, the enhancement processing includes a plurality of sub-transformation processes; and the adjusting of the initial enhancement parameter based on the feature information to obtain a current enhancement parameter includes: respectively obtaining the current enhancement parameter corresponding to each of the sub-transformation processes based on the feature information.

[0030] Through the above technical solution, the current enhancement parameters corresponding to various sub-transformation processes are adjusted, the automation, adaptability and accuracy are improved, the need for manual intervention is reduced, and the efficiency and reliability of subsequent processing of the case report picture are improved.

[0031] Optionally, the receiving of the case report picture uploaded by the current user terminal includes:

[0032] In response to an operation of calling a camera of the current user terminal by the current user, environment information and historical correction information corresponding to the camera are obtained, the environment information includes brightness information and a shooting angle, and the historical correction information is a parameter representing a correction condition of a case report picture shot by the camera;

[0033] Based on the environment information and the historical correction information, guide information corresponding to the camera is obtained, and the guide information is sent to the current user terminal, so that the current user terminal guides the current user to shoot the case report picture based on the guide information.

[0034] Through the above technical solution, when the camera is called to upload the case report picture, the operation of the current user for collecting the case report picture through the camera is guided by the guide information, so as to reduce the possibility that the collected case report picture needs to be corrected and improve the accuracy of reading the case report picture.

[0035] Optionally, the obtaining of the number threshold corresponding to the current user includes:

[0036] If the total number of uploads is one, that is, the case report picture is uploaded for the first time, the initial number threshold is taken as the number threshold corresponding to the current user;

[0037] If the total number of uploads is greater than one, and the current user manually sets the requirement for replacing the clinical expert, i.e., the clinical expert selected in this analysis is different from the corresponding clinical expert in the historical operation, the initial number threshold is corrected by the first number correction amount, and the number threshold obtained by adding the first number correction amount to the initial number threshold is taken as the number threshold corresponding to the current user;

[0038] If the number of uploads is greater than the initial number threshold, and the reply information allowing multiple uploads sent by the current user terminal is received, the initial number threshold is corrected by the second number correction amount, and the number threshold obtained by adding the second number correction amount to the initial number threshold is taken as the number threshold corresponding to the current user.

[0039] Through the above technical solutions, the adaptability of the number threshold to the actual situation of the current user is improved, the automation, adaptability and accuracy are improved, and the need for manual intervention is reduced.

[0040] In a second aspect of the present application, a cervical lesion related examination result analysis device is provided.

[0041] A cervical lesion related examination result analysis device comprises:

[0042] A receiving and reading module is configured to receive a case report picture uploaded by a current user terminal, and read case data in the case report picture.

[0043] A first obtaining module is configured to obtain prediction result information corresponding to the case data based on a current prediction model, and select optimal expert information based on the prediction result information.

[0044] A pushing and receiving module is configured to push the case data and the prediction result information to an expert terminal corresponding to the optimal expert information, and receive a final clinical analysis result fed back by the expert terminal.

[0045] A sending module is configured to send the final clinical analysis result to a current user terminal corresponding to the current user.

[0046] In a third aspect of the present application, an electronic device is provided, comprising a processor and a memory, wherein the processor is coupled to the memory.

[0047] The processor is configured to execute a computer program stored in the memory, so that the electronic device executes the method of any one of the first aspect.

[0048] In a fourth aspect of the present application, a computer readable storage medium is provided, comprising instructions which, when executed on a computer, cause the computer to execute the method of any one of the first aspect.

[0049] It should be understood that the content described in the summary section is not intended to define key or essential features of embodiments of the application or to limit the scope of the application. Other features of the application will be apparent from the description that follows. BRIEF DESCRIPTION OF DRAWINGS

[0050] The objects, features and advantages of the embodiments of the present application will be more clearly understood from the following detailed description taken in conjunction with the accompanying drawings. In the drawings:

[0051] Fig. 1 is a flow diagram of a method for analyzing results of examinations related to cervical lesions according to an embodiment of the present application;

[0052] Fig. 2 is a block diagram of a device for analyzing results of examinations related to cervical lesions according to an embodiment of the present application;

[0053] Fig. 3 is a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the skilled in the art better understand the technical scheme in one or more embodiments of the present specification, the technical scheme in one or more embodiments of the present specification will be described clearly and completely below in conjunction with the drawings in one or more embodiments of the present specification. Obviously, the described embodiments are only a part of the embodiments of the present specification, not all the embodiments. Based on one or more embodiments of the present specification, all other embodiments obtained by those skilled in the art without creative labor should belong to the protection scope of the present document.

[0055] It should be noted that the embodiments described by the present application are only to more clearly illustrate the technical scheme of the embodiments of the present application, and do not constitute a limitation on the technical scheme provided by the embodiments of the present application.

[0056] The embodiments of the present application provide a method for analyzing results of examinations related to cervical lesions. The method can be executed by a device, which can be a server or a terminal device. The server can be a physical server, a server cluster composed of multiple physical servers, a distributed system, or a cloud server providing cloud computing services. The terminal device can be a smartphone, a tablet computer, a desktop computer, etc., but is not limited thereto.

[0057] As shown in Fig. 1, a method for analyzing results of examinations related to cervical lesions is executed by an electronic device. The main process of the method is described as follows (steps S111-S114):

[0058] Step S111: receiving a case report picture uploaded by the current user terminal; reading case data in the case report picture.

[0059] In this embodiment, the case data can include personal basic information, cervical cytology test results, and / or human papillomavirus test results.

[0060] The case report picture can be an image uploaded through the local storage corresponding to the current user terminal, or an image uploaded after the case report is photographed by calling the camera of the current user terminal. The current user terminal sends the case report picture to the electronic device, and the electronic device processes the received case report picture through a preset reading algorithm to obtain the case data. It is easy to understand that the case report can be a cervical cytology report and / or a human papillomavirus test report and / or other cervical lesion related test report.

[0061] In this embodiment, the preset reading algorithm can be TableParser-OCR algorithm, and reading the case data in the case report picture based on the TableParser-OCR algorithm specifically includes the following processing:

[0062] Text region segmentation: the algorithm uses a text detection algorithm to detect the text region in the image. The text detection algorithm includes deep learning network DBNet or PSENet, etc. These algorithms are based on deep neural networks and can generate text boxes by learning the features and geometric shapes of text regions.

[0063] Text content recognition: for the detected text region, the TableParser-OCR algorithm uses a CRNN (Convolutional Recurrent Neural Network) model for character recognition. The CRNN model combines convolutional neural networks (CNN) and recurrent neural networks (RNN) and can perform end-to-end recognition on text regions.

[0064] Text content post-processing: after character recognition, the TableParser-OCR algorithm performs post-processing on the recognition results. This includes character-level error correction and text line sorting, etc. to improve recognition accuracy and readability of the results.

[0065] Multi-task learning: the TableParser-OCR algorithm uses multi-task learning method to process text region segmentation and text content recognition simultaneously. Through joint training of the two tasks, the overall recognition performance can be improved.

[0066] Model optimization and acceleration: TableParser-OCR algorithm uses a series of model optimization and acceleration techniques, such as model pruning, quantization and lightweight design, to improve the inference speed and deployment effect of the model.

[0067] TableParser-OCR algorithm realizes the detection and recognition of text regions in images through the joint work of text region segmentation and text content recognition, and can process text in various scenarios with high accuracy and robustness, and supports multi-language text recognition.

[0068] TableParser-OCR is a specific tool or method for extracting text data from case report pictures and outputting specific formats of text data, including text content, location information and confidence. The recognition results are converted into a data list format, with each piece of information consisting of a text or phrase sentence, a quadrilateral corner pixel coordinate (location information) and a confidence.

[0069] When using TableParser-OCR to extract text data from case report pictures, since the detection report formats corresponding to case report pictures provided by different suppliers may be different, it is necessary to establish a targeted information extraction template to ensure accurate extraction of the required information. Among them, the detection report refers to the written record of the results, conclusions and other contents of the detected sample issued by the detection agency after the detection process is completed, in this embodiment, the detection report is the picture report that needs to be processed by TableParser-OCR, that is, the original content before data recognition.

[0070] In this embodiment, the electronic device stores information extraction templates corresponding to the report formats of various detection reports, which are used to guide the extraction of specific types of information from text or pictures. After TableParser-OCR recognition, the information extraction template is used to analyze the recognized data to ensure that key information can be accurately captured. For example, when extracting the detection result of HPV16, different hospitals or detection agencies may use different terms or formats to identify this information, so it is necessary to establish an extraction template for each format, and the keywords or phrases defined in the information extraction template are extraction character markers, which are used to locate the required information in the data after TableParser-OCR recognition. For example, the extraction character marker in the "certain hospital detection report" is "human papillomavirus 16 subtype high risk", while the extraction character marker in the "certain hospital-human papilloma virus (HPV) gene detection report" is "HPV16".

[0071] In different formats of detection reports, targeted information extraction templates can be flexibly used to ensure accurate data extraction, so as to achieve comprehensive and accurate understanding of the report content.

[0072] In this embodiment, the electronic device includes a hierarchical template matching module, which includes a preferred matching template and multiple suboptimal matching templates. After completing the text conversion of the case report picture based on TableParser-OCR, the hierarchical template matching module can check the recognized case data according to the preset matching template, and the matching template defines the structure and format of the report. The electronic device organizes and formats the recognized case data according to the predetermined structure corresponding to the matching template according to the indication of the matching template, thereby generating structured report data.

[0073] The preferred matching template is a matching strategy with good efficiency and stability. In this embodiment, the preferred matching template includes key characters such as the position, order, font, size, etc. of each part of the case report picture, which have distinguishing significance. Exemplarily, the key characters can be "XX Hospital" or "Cytology Examination Report". The preferred matching template first detects whether the key characters are contained in the case data. If the key characters are detected, it is considered that the template matching of the structured report data is successful.

[0074] The suboptimal matching template is used to describe a more complex matching strategy that is suboptimal. If the key characters are not detected, it is considered that the template matching fails, and the suboptimal matching template can be used for detection. Exemplarily, the suboptimal matching template can be a template data matched by using a machine learning or deep learning algorithm. The suboptimal matching template can learn the features and patterns of the template through training the model, and can automatically adapt to new and unknown data. For example, a convolutional neural network (CNN) or a recurrent neural network (RNN) can be used to process the case data.

[0075] In this embodiment, if the matching strategies of multiple groups of matching templates all fail, that is, after trying multiple matching schemes, the template matching the case data still cannot be found, the content in the report cannot be accurately parsed or recognized according to the existing hierarchical template matching module. Exemplarily, the number of groups can be preset to 3 groups.

[0076] Step S112: obtaining the prediction result information corresponding to the case data based on the current prediction model, and selecting the optimal expert information based on the prediction result information.

[0077] Step S113: pushing the case data and the prediction result information to the expert terminal corresponding to the optimal expert information, and receiving the final clinical analysis result fed back by the expert terminal.

[0078] The prediction model is configured to output prediction result information corresponding to the case data. In this embodiment, the prediction model can adopt a supervised deep neural network algorithm as a technical model, and can output the probability of each category, so as to obtain the prediction result information. The sum of the probabilities is 1, and the category with the maximum probability is the classification result. The prediction result information includes the predicted value (0 or 1) and the corresponding confidence interval, i.e., the result of predicting the cervical lesion and the probability of occurrence.

[0079] The final clinical analysis result includes that the current user needs to receive further pathological examination and does not need to receive further pathological examination. After obtaining the prediction result information, the corresponding clinical expert of the expert terminal needs to further confirm the final clinical diagnosis result. Specifically, the clinical expert analyzes and judges the prediction result. If it is judged that the current user needs to receive further pathological examination, the final clinical diagnosis result is confirmed according to the pathological result of the further examination. If it is judged that the current user does not need to receive further pathological examination, the final clinical diagnosis result is defaulted to be negative according to the clinical experience, i.e., the current user does not have high-grade cervical intraepithelial neoplasia and cervical cancer.

[0080] The electronic device is preconfigured with expert information corresponding to a plurality of clinical experts. Therefore, in order to improve the accuracy and work efficiency of the judgment of the clinical experts, the optimal expert information needs to be selected, and the work of further confirming the final clinical diagnosis result is distributed to the expert terminal corresponding to the optimal expert information.

[0081] In this embodiment, the selection of the optimal expert information based on the prediction result information specifically includes the following processing: if the prediction result information is that cervical lesions occur or the probability of occurrence is high, the expert information of long service time and waiting judgment time not exceeding the first preset time threshold is selected as the optimal expert information; if the prediction result information is that cervical lesions do not occur or the probability of occurrence is low, the expert information of short service time or waiting judgment time not exceeding the second preset time threshold is selected as the optimal expert information, and the second preset time threshold is higher than the first preset time threshold.

[0082] The waiting judgment time is the waiting time length of the clinical expert corresponding to the expert information for further judgment of the current case data and the prediction result information. It is easy to understand that the less the waiting judgment time is, the higher the efficiency of further judgment of the current case data and the prediction result information is. The longer the clinical expert's service time is, the higher the accuracy of further confirming the final clinical diagnosis result is. The preliminary judgment is made through the prediction model, and the clinical experts are distributed according to the prediction result information, thereby improving the rationality of the use of medical resources.

[0083] It is easy to understand that the current user can manually select the expert information according to his own needs, and the expert information selected by the current user is regarded as the optimal expert information. After the current user manually selects the optimal expert information, there is no need to continue to perform the step of selecting the optimal expert information based on the prediction result information, that is, the priority of manually selecting the optimal expert information is higher than the priority of selection.

[0084] In this embodiment, the case data can also be pushed to the current user terminal corresponding to the current user. After pushing the case data to the current user terminal corresponding to the current user, the current user terminal is allowed to manually adjust or modify the information in the case data, or to select to abandon the currently entered case data result, and to re-perform the entire operation process of uploading the case report picture, data recognition and information extraction.

[0085] When the result of the case data automatically recognized by the electronic device does not meet the preset result requirement, for example, there is a character recognition error or an incomplete field statement recognition, the corresponding data field will be displayed as empty. At this time, the electronic device will prompt the current user to manually enter the data.

[0086] If there is data with positive test results in the recognized case data, the electronic device will automatically highlight these data results for the current user to carefully check and confirm.

[0087] The electronic device also processes the recognized case data according to the credibility determination of the algorithm model. When the credibility of the recognized text is lower than the set threshold, there may be a situation of word group recognition error. At this time, the electronic device can try to match with the preset typical value. If the matching is successful, these related results will be displayed in a highlighted form in the case data sent to the current user terminal for the current user to focus on checking. If the matching is unsuccessful, the electronic device will identify that these characters are not correctly recognized, and the current user needs to manually adjust or re-upload the picture for recognition.

[0088] Step S114: sending the final clinical analysis result to the current user terminal corresponding to the current user.

[0089] In this embodiment, the final clinical analysis result is sent to the current user terminal, which is convenient for the current user to check, and whether to perform further pathological examination is selected according to the final clinical analysis result, thereby improving the rationality of the use of medical resources.

[0090] As an optional implementation manner of this embodiment, the current user can be bound with an associated user. The associated user can be a relative of the current user, and the current user terminal is a terminal operated by the associated user. Therefore, the final clinical analysis result can be sent to the associated user, thereby reducing the possibility that the current user is inconvenient to operate and the final clinical analysis result cannot be checked.

[0091] In this embodiment, the case report picture may have an uploading error, in order to reduce the influence of the case report picture uploading error and improve the accuracy of the prediction result information, the following processing is further included before step S112: obtaining a current report type corresponding to the case report picture; if the current report type is a preset report type related to cervical lesion, obtaining a total uploading times corresponding to the case report picture; obtaining a times threshold corresponding to the current user, and determining whether the total uploading times reaches the times threshold; if yes, sending a first prompt information to the current user terminal; if no, performing the step of obtaining the prediction result information corresponding to the case data based on the current prediction model.

[0092] The electronic device stores difference data between case data of different report types, after reading the case data in the case report picture, the current report type corresponding to the case report picture can be determined according to the case data.

[0093] The preset report type includes a cervical cytology report type, a human papilloma virus report type and other cervical lesion related examination report types, the current report type is compared with the preset report type, if the current report type exists in the preset report type, the current report type is the preset report type related to cervical lesion, otherwise, the current report type is not the preset report type.

[0094] The total uploading times is the sum of the total times of uploading the case report picture in the historical operation and the current times, the electronic device stores an initial times threshold, a first times correction amount and a second times correction amount, for example, the initial times threshold can be three times, the first times correction amount is two times, and the second times correction amount is five times.

[0095] The obtaining of the times threshold corresponding to the current user specifically includes the following processing:

[0096] If the total uploading times is one, that is, the case report picture is uploaded for the first time, the initial times threshold is taken as the times threshold corresponding to the current user;

[0097] If the total uploading times is greater than one, and the current user manually sets the demand for replacing the clinical expert, that is, the clinical expert selected in this analysis is different from the corresponding clinical expert in the historical operation, the initial times threshold is corrected by the first times correction amount, and the times threshold obtained by adding the first times correction amount and the initial times threshold is taken as the times threshold corresponding to the current user;

[0098] If the uploading times is greater than the initial times threshold, and the reply information of allowing multiple uploading sent by the current user terminal is received, the initial times threshold is corrected by the second times correction amount, and the times threshold obtained by adding the second times correction amount and the initial times threshold is taken as the times threshold corresponding to the current user.

[0099] In the embodiment, since different clinical experts may report different final clinical analysis results for the same case report picture, in order to meet the requirement of the user for the pathological report to be judged by different clinical experts, the number threshold obtained by adding the first number of correction to the initial number threshold is taken as the number threshold corresponding to the current user.

[0100] The electronic device performs user profiling on the current user according to the received reply information allowing multiple uploads, and labels the current user as a user requiring multiple analyses, so as to realize adaptive adjustment of the corresponding number of requirements according to user habits.

[0101] In the embodiment, the first prompt information can be prompt information of a preset format and preset content, and the first prompt information can be in the form of text, voice, etc. The current user is inquired through the first prompt information to determine whether the case report picture is uploaded by mistake, so as to reduce the influence of the case report picture uploaded by mistake and improve the accuracy of the prediction result information.

[0102] In the embodiment, after obtaining the current report type corresponding to the case report picture, the following processing is further included: obtaining the corresponding current model type based on the current report type; and taking the prediction model of the current model type as the current prediction model.

[0103] The electronic device stores model types and prediction models corresponding to different report types. Since the current user may have one or more of cervical cytology reports, human papilloma virus detection reports, and other cervical lesion related examination reports, when the user uploads one or more different case report pictures, the electronic device will select the corresponding model type according to the above one or more report types, and predict the case data through the prediction model corresponding to the model type, so as to improve the adaptability to different report type combinations possessed by the user.

[0104] In the embodiment, the preset optimal type is stored in the electronic device, and the preset optimal type can be a cervical cytology report type and a human papilloma virus report type. Before the reading of the case data in the case report picture in step S111, the following processing is further included: if the current report type is not the preset optimal type, obtaining feature information in the case report picture; adjusting the initial enhancement parameter based on the feature information to obtain a current enhancement parameter; and performing enhancement processing on the case report picture based on the current enhancement parameter.

[0105] In this embodiment, the feature information includes image edges, textures, shapes and sizes. By performing enhancement processing on the case report picture, various changes that may occur in the actual shooting process of the case report picture can be simulated, thereby achieving data expansion. When the current report type is not the best preset type, data expansion is performed to improve the accuracy of subsequent case data reading and reduce the impact of an undesirable report type on case data reading.

[0106] The enhancement processing includes various sub-transformation processing such as rotation, scaling, translation, flipping and elastic transformation. The initial enhancement parameters are adjusted based on the feature information to obtain the current enhancement parameters, which specifically include the following processing: the current enhancement parameters corresponding to various sub-transformation processing are obtained based on the feature information respectively.

[0107] The electronic device is provided with threshold values corresponding to various sub-transformation processing. These threshold values can be obtained based on experience, statistical data or prediction results of a machine learning model. The feature information is compared with the corresponding threshold values to identify which feature information is consistent with the expected target and which feature information needs to be processed by the corresponding sub-transformation processing.

[0108] For example, when the structure or lesion site in the case report picture deviates from the image center, the rotation angle is calculated according to the corresponding feature information and the position standard to make the lesion site closer to the image center; when the contrast of the case report picture is insufficient, the contrast needs to be increased; and when the case report picture is blurred, sharpening or denoising technology needs to be applied. According to the determined sub-transformation processing, the difference between the feature information and the threshold value is used to adjust the initial enhancement parameters to obtain the current enhancement parameters.

[0109] According to the feature information and the set threshold values, the current enhancement parameters corresponding to various sub-transformation processing are adjusted to improve the degree of automation, adaptability and accuracy, reduce the need for manual intervention, and improve the efficiency and reliability of subsequent processing of the case report picture.

[0110] In this embodiment, if the current user needs to call the camera to upload the case report picture, the case report picture uploaded by the current user terminal includes the following processing: in response to the operation of the current user calling the camera of the current user terminal, the environment information and the historical correction information corresponding to the camera are obtained, the environment information includes brightness information and shooting angle, and the historical correction information is a parameter representing the correction of the case report picture shot by the camera; based on the environment information and the historical correction information, the guide information corresponding to the camera is obtained and sent to the current user terminal to guide the current user to shoot the case report picture based on the guide information.

[0111] The electronic device stores historical correction information corresponding to a user, the historical correction information including correction content in historical correction operations, the historical correction operations being manual corrections performed by the user on case data when the case data in a read case report picture does not match case data in a case report, the matching being determined by comparing whether the case data content of both is correct and whether the case data content is missing. For example, when the user terminal camera has a hardware defect such as being unable to normally capture picture content in a certain direction, the case data content is missing, and the picture content in the direction needs to be corrected in the historical correction operations to make the case data in the read case report picture consistent with the case data in the case report. Therefore, the same correction reason that exceeds a preset number threshold in the historical correction information can be used as a potential risk.

[0112] The environmental information is an environmental parameter of a current user terminal, wherein the brightness information can be collected by a brightness sensor of the current user terminal, and the shooting angle can be collected by a gyroscope of the current user terminal. In this embodiment, the electronic device stores a target brightness and a first camera angle.

[0113] In this embodiment, the guidance information includes brightness guidance and angle guidance. The electronic device generates the brightness guidance according to a difference between the target brightness and the current brightness information. The current user terminal has a fill light. After receiving the brightness guidance, the current user terminal controls the fill light to be turned on, and adjusts the environmental brightness where the current user terminal is located to the target brightness.

[0114] For example, the angle guidance can be a text display of the adjustment direction and the angle value on the display screen of the current user terminal.

[0115] If the same correction reason that exceeds the preset number threshold in the historical correction information is not a camera angle failure reason, the electronic device generates the angle guidance according to a difference between the first camera angle and the current shooting angle.

[0116] If the same correction reason that exceeds the preset number threshold in the historical correction information is a camera angle failure reason, the difference between the current camera angle and the ideal shooting angle can be determined based on the alignment between the case report picture and the corrected picture in the historical correction information. The angle correction amount is determined based on the deviation amount. The first camera angle is corrected based on the angle correction amount to obtain a second camera angle. The electronic device generates the angle guidance according to a difference between the second camera angle and the current shooting angle.

[0117] In this embodiment, the angle correction amount is intended to adjust the shooting angle of the camera from the problematic angle range to a more suitable angle to optimize the shooting effect. The angle correction amount can be a specific numerical value. In some cases, it is desirable to shoot case report pictures from different angles to obtain more comprehensive information, and there are multiple suitable correction angles. Therefore, the electronic device can calculate one or more angle correction amounts for selection according to the needs. When calculating the correction angle, the electronic device also needs to consider the shooting target, which is a preset content in the electronic device. For example, the shooting target is to shoot a specific data area in the case report, and the electronic device preferentially selects the correction angle that can clearly shoot this part as the angle correction amount.

[0118] When shooting paper case reports, the physical form of the case report (such as folds, bends, or wrinkles) may cause the image quality of the case report picture to decrease, thereby affecting the accuracy of recognition. Therefore, before reading the case data in the case report picture, the case report picture also needs to be checked to ensure that the case report picture quality meets the requirements.

[0119] In this embodiment, the checking method of the case report picture quality can be to check the quadrilateral quality of the segmentation area. The segmentation area is the position area calibrated by the TableParser-OCR algorithm for each character or phrase in the recognition process. If the quadrilateral of the segmentation area is a rectangle, and the midlines of the long opposite sides of different segmentation quadrilaterals have consistency, it means that the case report picture is relatively flat and does not have serious wrinkles or quality problems. On the contrary, if the consistency is poor, it means that the case report picture has wrinkles or poor image quality. When the included angle of the long opposite sides of the quadrilateral of the segmentation area is close to zero degrees, it can be determined that the quadrilateral is a rectangle.

[0120] In this embodiment, after reading the case data in the case report picture, the following processing is also included: obtaining the treatment plan information corresponding to the current user, and obtaining the target data corresponding to the current treatment node based on the treatment plan information; comparing the case data with the target data, obtaining the second prompt information based on the comparison result, and sending the second prompt information to the expert terminal.

[0121] The electronic device stores different user corresponding treatment plan information, and the treatment plan information includes multiple treatment nodes, treatment schemes corresponding to the treatment nodes, and target data corresponding to the treatment nodes. The target data is each item of data in the report obtained by performing a cervical lesion related examination again after treatment by the treatment scheme corresponding to the previous treatment node.

[0122] The case data obtained by the reading is the data after the treatment of the treatment scheme corresponding to the last treatment node. By comparing the case data with the target data, the deviation value of the case data and the target data can be obtained. When the deviation value exceeds the corresponding preset deviation value threshold, the second prompt information is generated. The second prompt information can be prompt information in a preset format and preset content. The second prompt information can be in the form of text, voice, etc. The second prompt information is used to remind the clinical expert corresponding to the expert terminal, so that the clinical expert can adjust the subsequent treatment scheme in a timely manner according to the second prompt information.

[0123] In this embodiment, step S114 specifically includes the following processing: based on the final clinical analysis result, report layout information is obtained, the report layout information including font, font size and character spacing; based on the report layout information, an analysis report is generated, and the analysis report is sent to the current user terminal.

[0124] The electronic device collects information such as the number of words and the size of the table of the final clinical analysis result. The electronic device stores a corresponding relationship between the information corresponding to the final clinical analysis result and the report layout information. The report layout information such as font, font size and character spacing is adaptively adjusted, so that the analysis report is clearer and easier to read.

[0125] Based on the same technical concept, after introducing the method in the embodiments of the application, next, with reference to FIG. 2, a cervical lesion related examination result analysis device 200 in the embodiments of the application is described. The device includes:

[0126] The receiving and reading module 201 is configured to receive a case report picture uploaded by a current user terminal, and read case data in the case report picture.

[0127] The first obtaining module 202 is configured to obtain prediction result information corresponding to the case data based on a current prediction model, and select optimal expert information based on the prediction result information.

[0128] The push and receive module 203 is configured to push the case data and the prediction result information to an expert terminal corresponding to the optimal expert information, and receive a final clinical analysis result fed back by the expert terminal.

[0129] The sending module 204 is configured to send the final clinical analysis result to a current user terminal corresponding to the current user.

[0130] In an optional embodiment, before the first obtaining module 202, the device further includes:

[0131] The second obtaining module is configured to obtain a current report type corresponding to the case report picture.

[0132] The third obtaining module is configured to, when the current report type is a preset report type related to cervical lesions, obtain a total number of times of uploading of the case report picture;

[0133] The fourth obtaining module is configured to obtain a number threshold corresponding to the current user, and determine whether the total number of times of uploading reaches the number threshold. If yes, the first prompt information is sent to the current user terminal. If no, the step of obtaining the prediction result information corresponding to the case data based on the current prediction model is executed.

[0134] In an optional implementation, after the second obtaining module, the apparatus further comprises:

[0135] The fifth obtaining module is configured to obtain a current model type corresponding to the current report type.

[0136] The module is configured to use a prediction model of the current model type as the current prediction model.

[0137] In an optional implementation, before reading the case data in the case report picture, the apparatus further comprises:

[0138] The sixth obtaining module is configured to, when the current report type is not a preset optimal type, obtain feature information in the case report picture.

[0139] The adjusting obtaining module is configured to adjust an initial enhancement parameter based on the feature information to obtain a current enhancement parameter.

[0140] The enhancement processing module is configured to perform enhancement processing on the case report picture based on the current enhancement parameter.

[0141] In an optional implementation, the adjusting obtaining module comprises:

[0142] The obtaining parameter submodule is configured to obtain a current enhancement parameter corresponding to each of the sub-transformation processes based on the feature information.

[0143] In an optional implementation, the receiving submodule comprises:

[0144] The response obtaining submodule is configured to, in response to an operation of invoking a camera of the current user terminal by the current user, obtain environment information and historical correction information corresponding to the camera, the environment information comprising brightness information and a shooting angle, and the historical correction information being a parameter representing a correction condition of a case report picture photographed by the camera.

[0145] The obtaining and sending sub-module is configured to obtain the guidance information corresponding to the camera based on the environmental information and the historical correction information, and send the guidance information to the current user terminal, so that the current user terminal guides the current user to take the case report picture based on the guidance information.

[0146] In an optional implementation, the fourth obtaining module comprises:

[0147] If the total uploading times is one, i.e., the case report picture is uploaded for the first time, the initial times threshold is taken as the times threshold corresponding to the current user.

[0148] If the total uploading times is more than one, and the current user manually sets the requirement for changing the clinical expert, i.e., the selected clinical expert in the current analysis is different from the corresponding clinical expert in the historical operation, the initial times threshold is corrected by the first times correction amount, and the times threshold obtained by adding the first times correction amount and the initial times threshold is taken as the times threshold corresponding to the current user.

[0149] If the uploading times is more than the initial times threshold, and the reply information allowing multiple uploading sent by the current user terminal is received, the initial times threshold is corrected by the second times correction amount, and the times threshold obtained by adding the second times correction amount and the initial times threshold is taken as the times threshold corresponding to the current user.

[0150] In an example, the modules in any of the above apparatuses can be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASICs), or one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0151] For another example, when the modules in the apparatus can be implemented in the form of a processing element scheduler, the processing element can be a general purpose processor, such as a central processing unit (CPU) or other processor that can invoke programs. For another example, these modules can be integrated together to be implemented in the form of a system-on-a-chip (SOC).

[0152] Various objects in the present application may be named, which may appear in the present application. It can be understood that these specific names do not constitute a limitation on the related objects, and the names can be changed according to the scene, context or use habits and other factors. The technical meaning of the technical terms in the present application should be mainly determined from the function and technical effect embodied / executed in the technical scheme.

[0153] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working process of the system, device and module described above can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0154] Those skilled in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical scheme. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0155] Based on the same technical concept, after introducing the method in the embodiments of the present application, next, referring to FIG. 3, an electronic device in the embodiments of the present application is described, which includes a processor 301 and a memory 302, and can further include one or more of an information input / output I / O interface 303, a communication component 304 and a communication bus 305.

[0156] The processor 301 is configured to control overall operations of the electronic device 300 to complete all or part of the steps of the cervical lesion related examination result analysis method described above. The memory 302 is configured to store various types of data to support operations of the electronic device 300. The data can include, for example, instructions for any application or method operating on the electronic device 300, and application-related data. The memory 302 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as one or more of a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk, or an optical disk.

[0157] The I / O interface 303 provides an interface between the processor 301 and other interface modules, which can be a keyboard, a mouse, a button, and the like. The buttons can be virtual buttons or physical buttons. The communication component 304 is configured to test wired or wireless communication between the electronic device 300 and other devices. The wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G, or 4G, or a combination of one or more of them, so the corresponding communication component 304 can include a Wi-Fi component, a Bluetooth component, and an NFC component.

[0158] The communication bus 305 can include a path for transmitting information between the above-mentioned components. The communication bus 305 can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus 305 can be divided into an address bus, a data bus, a control bus, etc.

[0159] The electronic device 300 can be implemented with one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic elements for performing the cervical lesion related examination result analysis method according to the above-described embodiments.

[0160] The electronic device 300 can include, but is not limited to, a mobile terminal such as a digital broadcasting receiver, a PDA (Personal Digital Assistant), a PMP (Portable Multimedia Player), and the like, and a stationary terminal such as a digital TV, a desktop computer, and the like, and can also be a server or the like.

[0161] Based on the same technical concept, after the electronic device in the embodiments of the present application is introduced, next, a computer readable storage medium in the embodiments of the present application is described, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the cervical lesion related examination result analysis method described above.

[0162] The computer readable storage medium can include a U disk, a mobile hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0163] The term "comprising" or "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or device.

[0164] The above description is only the preferred embodiment of the present application and the explanation of the applied technical principles. It should be understood by those skilled in the art that the inventive scope of the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by the combinations of the above technical features or their equivalent features without departing from the inventive concept. For example, the technical solutions formed by the mutual replacement of the above features and the technical features with similar functions disclosed in the present application (but not limited to) without departing from the inventive concept.

Claims

1. A method of analyzing results of a cervical lesion-related examination, characterized by, The method comprises the following steps: receiving a case report picture uploaded by a current user terminal, and reading case data in the case report picture; the case data comprises personal basic information and cervical cytology examination results; obtaining prediction result information corresponding to the case data based on a current prediction model, and selecting optimal expert information based on the prediction result information; the prediction result information comprises a calculated prediction value and a corresponding confidence interval, i.e. a result of predicting cervical lesions and a probability of occurrence; pushing the case data and the prediction result information to an expert terminal corresponding to the optimal expert information, and receiving a final clinical analysis result fed back by the expert terminal; sending the final clinical analysis result to a current user terminal corresponding to the current user.

2. The method of claim 1, wherein the method comprises: Before the step of obtaining the prediction result information corresponding to the case data based on the current prediction model, the method further comprises the following steps: obtaining a current report type corresponding to the case report picture; if the current report type is a preset report type related to cervical lesions, obtaining a total number of uploads corresponding to the case report picture; obtaining a number threshold corresponding to the current user, and determining whether the total number of uploads reaches the number threshold; if yes, sending a first prompt information to the current user terminal; if no, performing the step of obtaining the prediction result information corresponding to the case data based on the current prediction model.

3. The method according to claim 2, wherein, After the step of obtaining the current report type corresponding to the case report picture, the method further comprises the following steps: based on the current report type, obtaining a current model type corresponding thereto; taking a prediction model of the current model type as the current prediction model.

4. The method according to claim 3, wherein, Before the step of reading the case data in the case report picture, the method further comprises the following steps: if the current report type is not a preset optimal type, obtaining feature information in the case report picture; adjusting an initial enhancement parameter based on the feature information to obtain a current enhancement parameter; based on the current enhancement parameter, performing enhancement processing on the case report picture.

5. The method of claim 4, wherein the method further comprises: The enhancement processing comprises a plurality of sub-transformation processes; the step of adjusting the initial enhancement parameter based on the feature information to obtain the current enhancement parameter comprises the following steps of respectively obtaining a current enhancement parameter corresponding to each of the sub-transformation processes based on the feature information.

6. The method of claim 1, wherein the method further comprises: The step of receiving the case report picture uploaded by the current user terminal comprises the following steps: in response to an operation of calling a camera of the current user terminal by a current user, obtaining environment information and historical correction information corresponding to the camera, wherein the environment information comprises brightness information and a shooting angle, and the historical correction information is a parameter representing a correction condition of a case report picture photographed by the camera; based on the environment information and the historical correction information, obtaining guide information corresponding to the camera, and sending the guide information to the current user terminal, so that the current user terminal guides the current user to photograph the case report picture based on the guide information.

7. The method of claim 2, wherein the method further comprises: The step of obtaining the number threshold corresponding to the current user comprises the following steps: if the total number of uploads is one, i.e. the case report picture is uploaded for the first time, taking an initial number threshold as the number threshold corresponding to the current user; If the total number of uploads is greater than one, and the current user manually sets the requirement for changing the clinical expert, i.e., the clinical expert selected in this analysis is different from the corresponding clinical expert in the historical operation, the initial number threshold is corrected by the first number correction amount, and the number threshold obtained by adding the first number correction amount to the initial number threshold is taken as the number threshold corresponding to the current user; If the number of uploads is greater than the initial number threshold, and the reply information allowing multiple uploads sent by the current user terminal is received, the initial number threshold is corrected by the second number correction amount, and the number threshold obtained by adding the second number correction amount to the initial number threshold is taken as the number threshold corresponding to the current user.

8. A device for analyzing results of examinations related to cervical lesions, characterized by Comprise: The receiving reading module is used for receiving the case report picture uploaded by the current user terminal, and reading the case data in the case report picture; The case data includes personal basic situation and cervical cytology examination result; The first acquisition module is used for acquiring the prediction result information corresponding to the case data based on the current prediction model, and selecting the optimal expert information based on the prediction result information; the prediction result information includes the calculated prediction value and the corresponding confidence interval, i.e., the result and the probability of predicting cervical lesions; The push receiving module is used for pushing the case data and the prediction result information to the expert terminal corresponding to the optimal expert information, and receiving the final clinical analysis result fed back by the expert terminal; The sending module is used for sending the final clinical analysis result to the current user terminal corresponding to the current user.

9. An electronic device, comprising: Comprise a processor and a memory, the processor is coupled with the memory; The processor is used for executing the computer program stored in the memory, so that the electronic equipment executes the method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It includes instructions that, when executed on a computer, cause the computer to perform the method of any one of claims 1-7.

Images