Medical auxiliary auditing method and system for image recognition

Abstract

The invention discloses a medical auxiliary auditing method and system for image recognition, and relates to the technical field of image processing, and the method comprises the steps: interacting with medical imaging equipment connected with a hospital information system through an interface, and exporting an image frame sequence of a target user in a preset window; traversing the image frame sequence to perform cross-frame feature associated memory recognition update to obtain a target associated memory unit; taking the target user queuing grade as target auxiliary auditing information; interacting a queuing management module of the hospital information system to obtain a target matching department; and in combination with the target auxiliary auditing information, adding the queuing application of the target user into a queue to be queued for optimization to obtain a target queuing queue. The technical problems of low intelligent triage efficiency and poor auxiliary auditing reliability of a hospital information system caused by deep mining of image data of a user in the prior art are solved, and the technical effects of improving the accuracy of auxiliary auditing materials and improving the triage queuing efficiency and quality are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to a medical auxiliary review method and system for image recognition. Background Art

[0002] Currently, hospital information systems (HIS) usually have functions such as patient information management, registration, queuing, diagnosis, and expense settlement. Some systems already support data docking with imaging acquisition devices to achieve the archiving and viewing of patient image information. However, most existing systems only stay at the level of image management or image browsing, and have not achieved auxiliary judgment and intelligent triage based on image content. Especially during the registration queuing process, it still mainly relies on users to independently select departments, lacking systematic analysis support for the degree of illness or structural abnormalities contained in the images, resulting in some patients with serious illnesses but not accurately identified not being able to receive timely and preferential treatment. Summary of the Invention

[0003] This application provides a medical auxiliary review method and system for image recognition, which is used to solve the technical problems in the prior art that there is a lack of in-depth mining of users' image data, resulting in low efficiency of intelligent triage in hospital information systems and poor reliability of auxiliary review.

[0004] In view of the above problems, this application provides a medical auxiliary review method and system for image recognition.

[0005] In the first aspect of this application, a medical auxiliary review method for image recognition is provided. The method includes: Interact with a medical imaging device connected to the hospital information system through an interface to export an image frame sequence of a target user within a preset window; traverse the image frame sequence for cross-frame feature correlation memory recognition update to obtain a target correlation memory unit; call a target user queuing level recognizer to extract information from the target correlation memory unit to determine the target user queuing level, and use the target user queuing level as target auxiliary review information; interact with the queuing management module of the hospital information system, perform department matching based on the target auxiliary review information to obtain a target matching department; perform a search for the to-be-queued queue based on the target matching department, and combine the target auxiliary review information to add the registration application of the target user into the to-be-queued queue for optimization to obtain a target queuing queue.

[0006] Preferably, based on the target matching department, a search for the queue to be numbered is performed, and in combination with the target auxiliary review information, the numbering application of the target user is added to the queue to be numbered for optimization, obtaining a target numbered queue, including: traversing the queue to be numbered to extract auxiliary review information for the queue to be numbered, obtaining a queue of auxiliary review information for the queue to be numbered; optimizing the target auxiliary review information and the queue of auxiliary review information for the queue to be numbered in descending order of the numbering level, obtaining the target numbered queue.

[0007] Preferably, traversing the image frame sequence for cross-frame feature correlation memory recognition update, obtaining a target correlation memory unit, including: Performing feature extraction on the image frame sequence respectively according to a preset image feature set, obtaining an image frame feature vector sequence, wherein the preset image feature set includes structural texture features, edge information, shape features, and gray-scale density features; Performing cross-frame feature correlation memory update based on the image frame feature vector sequence, obtaining a target correlation memory unit.

[0008] Preferably, performing feature extraction on the image frame sequence respectively according to a preset image feature set, obtaining an image frame feature vector sequence, including: obtaining a sample image frame set, respectively extracting the structural texture features, edge information, shape features, and gray-scale density features of each sample image frame in the sample image frame set according to the preset image feature set, filling them into an initially empty vector, obtaining a sample image frame feature vector set; obtaining a feature extractor, wherein the feature extractor is constructed according to the mapping relationship between the sample image frame set and the sample image frame feature vector set; using the feature extractor to perform feature extraction on the image frame sequence, obtaining the image frame feature vector sequence.

[0009] Preferably, performing cross-frame feature correlation memory update based on the image frame feature vector sequence, obtaining a target correlation memory unit, including: extracting a first image frame feature vector and a second image frame feature vector from the image frame feature vector sequence; performing cross-frame feature correlation analysis on the first image frame feature vector and the second image frame feature vector, obtaining a second correlated image frame feature vector, adding the first image frame feature vector and the second correlated image frame feature vector to a first correlation memory unit; extracting a third image frame feature vector from the image frame feature vector sequence again, using the first correlation memory unit to perform cross-frame feature correlation memory update on the third image frame feature vector, obtaining a second correlation memory unit; performing cross-frame feature correlation analysis on the remaining image frame feature vectors of the image frame feature vector sequence based on the second correlation memory unit in sequence, and updating the second correlation memory unit according to the analysis result, obtaining a target correlation memory unit.

[0010] Preferably, performing cross-frame feature correlation analysis on the first image frame feature vector and the second image frame feature vector to obtain a second correlated image frame feature vector, including: calculating the similarity of corresponding elements of the first image frame feature vector and the second image frame feature vector to obtain an element similarity set; performing normalization processing on the element similarity set and matrixifying the processed result to obtain a correlation analysis matrix; using the correlation analysis matrix to enhance the second image frame feature vector to obtain a second correlated image frame feature vector.

[0011] Preferably, the method further includes: using the correlation analysis matrix to perform weighted fusion on the second image frame feature vector to obtain a second enhanced image frame feature vector; using the weighted average method to fuse the second enhanced image frame feature vector and the second image frame feature vector to obtain the second correlated image frame feature vector.

[0012] Preferably, extracting a third image frame feature vector from the image frame feature vector sequence again, and using the first associative memory unit to perform cross-frame feature correlation memory update on the third image frame feature vector to obtain a second associative memory unit, including: using the first image frame feature vector in the first associative memory unit to perform cross-frame feature correlation analysis on the third image frame feature vector to obtain a first initial third correlated image frame feature vector; using the second correlated image frame feature vector in the first associative memory unit to perform cross-frame feature correlation analysis on the third image frame feature vector to obtain a second initial third correlated image frame feature vector; performing mean processing on the first initial third correlated image frame feature vector and the second initial third correlated image frame feature vector to obtain a third correlated image frame feature vector, and adding the third correlated image frame feature vector into the first associative memory unit to obtain a second associative memory unit.

[0013] Preferably, verifying the target auxiliary audit information and optimizing the network parameters of the target user queuing level identifier according to the verification result.

[0014] In a second aspect of the present application, a medical auxiliary audit system for image recognition is provided, and the system includes: An image frame sequence export module is used to interact with a medical imaging device connected to a hospital information system through an interface and export an image frame sequence of a target user within a preset window; a target associated memory unit acquisition module is used to traverse the image frame sequence for cross-frame feature associated memory recognition update to obtain a target associated memory unit; a target auxiliary review information acquisition module is used to call a target user queuing level recognizer to extract information from the target associated memory unit, determine the target user queuing level, and use the target user queuing level as target auxiliary review information; a target matching department acquisition module is used to interact with the queuing management module of the hospital information system, perform department matching based on the target auxiliary review information, and obtain a target matching department; a target queuing queue acquisition module is used to perform a search for a queuing queue to be processed based on the target matching department, combine the target auxiliary review information, add the queuing application of the target user to the queuing queue to be processed for optimization, and obtain a target queuing queue.

[0015] One or more technical solutions provided in this application have at least the following technical effects or advantages: In this application, by interacting with a medical imaging device connected to a hospital information system through an interface, an image frame sequence of a target user within a preset window is exported, then the image frame sequence is traversed for cross-frame feature associated memory recognition update to obtain a target associated memory unit, and then a target user queuing level recognizer is called to extract information from the target associated memory unit, determine the target user queuing level, use the target user queuing level as target auxiliary review information, then interact with the queuing management module of the hospital information system, perform department matching based on the target auxiliary review information, obtain a target matching department, and then perform a search for a queuing queue to be processed based on the target matching department, combine the target auxiliary review information, add the queuing application of the target user to the queuing queue to be processed for optimization, and obtain a target queuing queue. It achieves the technical effect of improving the accuracy of dynamic trend analysis of image recognition, thereby providing reliable auxiliary review information and providing reliable data support for optimizing the queuing queue. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0017] Figure 1 It is a schematic flowchart of a medical auxiliary review method for image recognition provided by an embodiment of this application; Figure 2 It is a schematic structural diagram of a medical auxiliary review system for image recognition provided by an embodiment of this application.

[0018] Explanation of the accompanying symbols: image frame sequence export module 11, target associated memory unit acquisition module 12, target auxiliary review information acquisition module 13, target matching department acquisition module 14, target queuing queue acquisition module 15. DETAILED DESCRIPTION

[0019] This application provides a medical auxiliary review method and system for image recognition, which is used to solve the technical problems in the existing technology of in-depth mining of user image data, resulting in low efficiency of intelligent triage in hospital information systems and poor reliability of auxiliary review.

[0020] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0021] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0022] Example 1, as Figure 1 As shown, the present application provides a medical auxiliary review method for image recognition, wherein the method includes: Step S100: The medical imaging device interacts with the hospital information system via an interface to export a sequence of image frames of the target user within a preset window; In one possible embodiment, the medical imaging device is an X-ray machine, CT (computed tomography), MRI (magnetic resonance imaging), ultrasound, or other device capable of acquiring structural images. The preset window is a time period predefined by a person skilled in the art, such as half a month, a month, or the like. The image frame sequence is obtained by exporting images captured by the medical imaging device within the preset window. The image frame sequence reflects the fluctuations in the target user's condition within the preset window.

[0023] By docking with a PACS system or an image acquisition server and a medical imaging device, an image frame sequence within a preset window is extracted and transmitted through reception to a hospital information system for subsequent analysis and processing, thereby providing a basis for analysis data for subsequent cross-frame feature correlation analysis. This step plays a role in controlling the data input source in the entire method and directly affects the context accuracy of feature extraction and the cross-frame correlation quality of memory update.

[0024] Step S200: Traverse the image frame sequence for cross-frame feature correlation memory recognition and update to obtain a target correlation memory unit; Further, traversing the image frame sequence for cross-frame feature correlation memory recognition and update to obtain a target correlation memory unit includes: Feature extraction is respectively performed on the image frame sequence according to a preset image feature set to obtain an image frame feature vector sequence, where the preset image feature set includes structural texture features, edge information, shape features, and gray-scale density features; Based on the image frame feature vector sequence, cross-frame feature correlation memory update is performed to obtain a target correlation memory unit.

[0025] Further, when feature extraction is respectively performed on the image frame sequence according to a preset image feature set to obtain an image frame feature vector sequence, step S200 of this application embodiment further includes: Obtain a sample image frame set, and based on the preset image feature set, respectively extract the structural texture features, edge information, shape features, and gray-scale density features of each sample image frame in the sample image frame set and fill them into an initially empty vector to obtain a sample image frame feature vector set; Obtain a feature extractor, where the feature extractor is constructed according to the mapping relationship between the sample image frame set and the sample image frame feature vector set; Use the feature extractor to perform feature extraction on the image frame sequence to obtain the image frame feature vector sequence.

[0026] In a possible embodiment, the preset image feature set refers to a set of multi-dimensional feature items defined in advance by those skilled in the art for identifying key contents of medical images, including structural texture features, edge information, shape features, and gray-scale density features. The sample image frame set is a labeled image set used to train the feature extractor. A feature vector is a structure that encodes multiple image features into a vector representation. The feature extractor is a trained model used to receive an image as input and output a multi-dimensional feature vector.

[0027] Obtain a set of sample image frames. Furthermore, those skilled in the art extract the structural texture features, edge information, shape features, and gray-scale density features of each sample image frame in the set of sample image frames according to the preset image feature set respectively, and fill the extraction results into an initially empty vector, so as to obtain a corresponding set of sample image frame feature vectors, where the vector representation is shown in Table 1.

[0028] Table 1 Sample Image Frame Feature Vector Table Preferably, use the set of sample image frames and the set of sample image frame feature vectors as the training data set, and use the training data set to perform supervised training on the framework constructed based on the convolutional neural network. During the training, learn the one-to-one mapping relationship between the sample image frames and the sample image frame feature vectors until the training converges, and obtain the trained feature extractor.

[0029] Furthermore, use the feature extractor to perform feature extraction on each image frame sequence one by one, and obtain the completed image frame feature vector sequence. Preferably, the image frame feature vector sequence reflects the representation of the image frame sequence.

[0030] By batch-processing the image frame sequence for feature extraction, the technical effect of providing data support for the subsequent process is achieved.

[0031] Further, perform cross-frame feature correlation memory update based on the image frame feature vector sequence to obtain the target correlation memory unit. The steps S200 of the embodiment of the present application further include: Extract the first image frame feature vector and the second image frame feature vector from the image frame feature vector sequence; Perform cross-frame feature correlation analysis on the first image frame feature vector and the second image frame feature vector to obtain the second associated image frame feature vector, and add the first image frame feature vector and the second associated image frame feature vector to the first associated memory unit; Extract the third image frame feature vector from the image frame feature vector sequence again, and use the first associated memory unit to perform cross-frame feature correlation memory update on the third image frame feature vector to obtain the second associated memory unit; Based on the second associated memory unit, perform cross-frame feature correlation analysis on the remaining image frame feature vectors of the image frame feature vector sequence in turn, and update the second associated memory unit according to the analysis results to obtain the target associated memory unit.

[0032] In a possible embodiment, the sequence of image frame feature vectors refers to an ordered set of features extracted from consecutive image frames by a feature extractor, which reflects information such as the structure, texture, shape, and grayscale of the image changing over time. Cross-frame feature correlation analysis refers to the process of comparing, fusing, or enhancing feature vectors from different image frames to identify common features or evolving features between frames. The associated image frame feature vector is an enhanced feature representation obtained by analysis between two frames, with stronger temporal semantics and context consistency. The associative memory unit: Essentially, it is a memory structure for storing and updating the cross-frame feature structure, used for progressively integrating the significant features between multiple frames.

[0033] In a possible embodiment, the sequence of image frame feature vectors is subjected to frame-by-frame correlation analysis through iterative operations to construct a memory structure. First, a first image frame feature vector and a second image frame feature vector are extracted from the sequence of image frame feature vectors. Through cross-frame feature correlation analysis, a second associated image frame feature vector is obtained. Among them, the second associated image frame feature vector contains the implicit trend of element changes from the first image frame feature vector to the second image frame feature vector. Furthermore, the first image frame feature vector and the second associated image frame feature vector are jointly added to the first associative memory unit. Among them, the first associative memory unit is used for subsequent cross-frame feature correlation analysis.

[0034] Next, a third associated image frame feature vector is extracted from the sequence of image frame feature vectors. Based on the existing content (the first image frame feature vector and the second associated image frame feature vector) in the foregoing first associative memory unit, the third associated image frame feature vector is further correlated and updated to output a new enhanced feature, and a second associative memory unit is extended and formed. This process continues iteratively, performing cross-frame analysis and fusion on the remaining image frames one by one, continuously updating the memory unit, and finally obtaining the target associative memory unit representing the associated expression of the entire image sequence.

[0035] Through inter-frame associated semantic mining and gradual knowledge accumulation, dynamic change trends can be identified, thereby enhancing the capabilities of the target user queuing level recognizer in temporal perception and context analysis, and further laying a solid foundation for generating high-confidence auxiliary review information.

[0036] Furthermore, when performing cross-frame feature correlation analysis on the first image frame feature vector and the second image frame feature vector to obtain the second associated image frame feature vector, step S200 of the embodiment of the present application further includes: Calculate the similarity of the corresponding elements of the first image frame feature vector and the second image frame feature vector to obtain a set of element similarities; Normalize the set of element similarities and matrixize the processed result to obtain an association analysis matrix; Enhance the second image frame feature vector using the correlation analysis matrix to obtain a second correlated image frame feature vector.

[0037] Furthermore, step S200 of the embodiment of the present application further includes: Perform weighted fusion on the second image frame feature vector using the correlation analysis matrix to obtain a second enhanced image frame feature vector; Fuse the second enhanced image frame feature vector and the second image frame feature vector using the weighted average method to obtain the second correlated image frame feature vector.

[0038] In an embodiment of the present application, use the cosine similarity calculation formula to calculate the similarity degree of the corresponding elements of the first image frame feature vector and the second image frame feature vector to obtain an element similarity set. Furthermore, use the min-max linear normalization processing method to normalize the element similarity set to obtain an element normalized similarity set. Furthermore, add the element normalized similarity set into an initially empty matrix to obtain a correlation analysis matrix. Wherein, the correlation analysis matrix reflects the correlation degree of different elements between the first image frame feature vector and the second image frame feature vector.

[0039] Furthermore, through matrix operations, perform weighted calculation on each dimension of features according to the weights in the correlation analysis matrix to obtain the second enhanced image frame feature vector. Wherein, the second enhanced image frame feature vector reflects the feature enhancement situation of the second image frame feature vector under the influence of the first image frame feature vector. Furthermore, perform weighted average on the second enhanced image frame feature vector and the second image frame feature vector to fuse the independent feature situation and the affected feature situation to obtain the second correlated image frame feature vector. By obtaining the second correlated image frame feature vector, the technical effects of improving the inter-frame semantic continuity and context consistency and providing higher-quality information input for subsequent construction of memory units are achieved.

[0040] Furthermore, extract a third image frame feature vector from the image frame feature vector sequence again, and use the first associated memory unit to perform cross-frame feature correlation memory update on the third image frame feature vector to obtain a second associated memory unit. Step S200 of the embodiment of the present application further includes: Perform cross-frame feature correlation analysis on the third image frame feature vector using the first image frame feature vector in the first associated memory unit to obtain a first initial third correlated image frame feature vector; Perform cross-frame feature correlation analysis on the third image frame feature vector using the second correlated image frame feature vector in the first associated memory unit to obtain a second initial third correlated image frame feature vector; Perform mean processing on the first initial third associated image frame feature vector and the second initial third associated image frame feature vector to obtain a third associated image frame feature vector, and add the third associated image frame feature vector into the first associated memory unit to obtain a second associated memory unit.

[0041] In an embodiment of the present application, based on the same principle as obtaining the second associated image frame feature vector, perform cross-frame feature correlation analysis on the third image frame feature vector based on the first image frame feature vector in the first associated memory unit to obtain a first initial third associated image frame feature vector. Among them, the first initial third associated image frame feature vector implies the correlation relationship between the third image frame feature vector and the first image frame feature vector. Similarly, perform cross-frame feature correlation analysis on the third image frame feature vector based on the second associated image frame feature vector in the first associated memory unit to obtain a second initial third associated image frame feature vector. Among them, the second initial third associated image frame feature vector implies the correlation relationship between the third image frame feature vector and the second associated image frame feature vector.

[0042] Furthermore, perform mean processing on the first initial third associated image frame feature vector and the second initial third associated image frame feature vector to obtain a third associated image frame feature vector. The third associated image frame feature vector reflects the implicit feature situation of the third image frame, and add the third associated image frame feature vector into the first associated memory unit to update it, obtaining a second associated memory unit.

[0043] Step S400: Invoke the target user queuing level recognizer to extract information from the target associated memory unit, determine the target user queuing level, and use the target user queuing level as the target auxiliary review information.

[0044] Further, verify the target auxiliary review information, and optimize the network parameters of the target user queuing level recognizer according to the verification result.

[0045] In a possible embodiment, obtain a plurality of sample target associated memory units and a plurality of sample target user queuing levels as training corpora. Use the training corpora to perform supervised training on a framework constructed based on a feedforward neural network, and verify the results output during training. When the verification passes, obtain the trained target user queuing level recognizer. Among them, the target user queuing level recognizer is a deep neural network with learning ability, which is used to perform discriminant analysis on the semantic features of the target associated memory unit and extract the target user queuing level.

[0046] Preferably, the training corpus is input into the framework to obtain multiple output target user queuing levels. The multiple output target user queuing levels are compared with the multiple sample target user queuing levels to obtain the success ratio of the comparison. When the ratio is greater than or equal to the ratio preset by those skilled in the art, the training is completed.

[0047] In a possible embodiment, when the ratio is less than or equal to the ratio preset by those skilled in the art, the network parameters of the target user queuing level recognizer are updated and adjusted according to the difference between the ratios until the requirements are met. The technical effect of improving the efficiency and reliability of the auxiliary review is achieved.

[0048] Step S400: Interact with the queuing management module of the hospital information system to perform department matching based on the target auxiliary review information to obtain the target matching department; Step S500: Based on the target matching department, perform a search for the queuing list to be arranged. Combine the target auxiliary review information, and add the queuing application of the target user into the queuing list to be arranged for optimization to obtain the target queuing list.

[0049] Furthermore, based on the target matching department, perform a search for the queuing list to be arranged. Combine the target auxiliary review information, and add the queuing application of the target user into the queuing list to be arranged for optimization to obtain the target queuing list. Step S500 of the embodiment of the present application further includes: Traverse the queuing list to be arranged to extract the auxiliary review information for the queuing list to be arranged to obtain the auxiliary review information queue for the queuing list to be arranged; Optimize the target auxiliary review information and the auxiliary review information queue for the queuing list to be arranged in descending order of the queuing level to obtain the target queuing list.

[0050] In a possible embodiment, according to the target auxiliary review information obtained after image recognition, it is mapped and interacted with the department information in the hospital queuing management module. Exemplarily, when the target auxiliary review information is the signs of cerebral hemorrhage detected in the image, it is subjected to semantic similarity matching with the set of department keywords preset by those skilled in the art, and the department keyword corresponding to the maximum semantic similarity is used as the matching department keyword. Furthermore, the department corresponding to the matching department keyword, such as the Department of Neurosurgery, is used as the target matching department. After obtaining the target matching department, it provides a clear direction for the next queuing optimization.

[0051] Perform a search in the waiting queue list of the target matching department to form a waiting queue. Further, extract the auxiliary review information of all users from the waiting queue and organize this information into a waiting queue for auxiliary review information, where each piece of waiting queue auxiliary review information reflects the queuing level of each user. Next, perform a level-priority comparison and sorting of the target user's auxiliary review information with the waiting queue for auxiliary review information. The sorting method is from high to low in terms of level. For example, users with a level of "urgent" are ranked before users with a level of "general" to achieve preferential guarantee of medical resources. The finally output new sequence is the target queuing queue. By using the auxiliary review results for optimizing the queuing logic, in-depth mining of the user's medical record images is realized, as well as the goal of improving triage efficiency.

[0052] In summary, the embodiments of the present application at least have the following technical effects: In the present application, a medical imaging device connected to the hospital information system through an interface is interacted with to export an image frame sequence of a target user within a preset window, and then the image frame sequence is traversed for cross-frame feature correlation memory recognition update to obtain a target correlation memory unit. Furthermore, a target user queuing level recognizer is called to extract information from the target correlation memory unit to determine the target user queuing level, and the target user queuing level is used as target auxiliary review information. Then, the queuing management module of the hospital information system is interacted with, and based on the target auxiliary review information, a target matching department is obtained. Furthermore, a search for the waiting queue is performed based on the target matching department, and in combination with the target auxiliary review information, the queuing application of the target user is added to the waiting queue for optimization to obtain a target queuing queue. The technical effect of improving the accuracy of dynamic trend analysis of image recognition is achieved, thereby providing reliable auxiliary review information and reliable data support for optimizing the queuing queue.

[0053] Embodiment 2, based on the same inventive concept as the medical auxiliary review method for image recognition in the foregoing embodiment, as Figure 2 shown, the present application provides a medical auxiliary review system for image recognition. The system in the embodiments of the present application and the method embodiments are based on the same inventive concept. Among them, the system includes: An image frame sequence export module 11, configured to interact with a medical imaging device connected to the hospital information system through an interface to export an image frame sequence of a target user within a preset window; A target correlation memory unit acquisition module 12, configured to traverse the image frame sequence for cross-frame feature correlation memory recognition update to obtain a target correlation memory unit; A target auxiliary review information acquisition module 13, configured to call a target user queuing level recognizer to extract information from the target correlation memory unit to determine the target user queuing level, and use the target user queuing level as target auxiliary review information; A target matching department obtaining module 14, which is used to interact with the queuing management module of the hospital information system, perform department matching based on the target auxiliary review information, and obtain a target matching department; A target queuing queue obtaining module 15, which is used to perform a search for a queuing queue to be processed based on the target matching department, combine the target auxiliary review information, and add the queuing application of the target user to the queuing queue to be processed for optimization, so as to obtain a target queuing queue.

[0054] Furthermore, the target queuing queue obtaining module 15 is used to perform the following steps: Traverse the queuing queue to be processed to extract auxiliary review information for the queuing queue to be processed, and obtain an auxiliary review information queue for the queuing queue to be processed; Optimize the target auxiliary review information and the auxiliary review information queue for the queuing queue to be processed in the order from the highest queuing level to the lowest, so as to obtain the target queuing queue.

[0055] Furthermore, the target associated memory unit obtaining module 12 is used to perform the following steps: Extract features from the image frame sequence respectively according to a preset image feature set, and obtain an image frame feature vector sequence, where the preset image feature set includes structural texture features, edge information, shape features, and gray-scale density features; Perform cross-frame feature associated memory update based on the image frame feature vector sequence, and obtain a target associated memory unit.

[0056] Furthermore, the target associated memory unit obtaining module 12 is used to perform the following steps: Obtain a sample image frame set, extract the structural texture features, edge information, shape features, and gray-scale density features of each sample image frame in the sample image frame set respectively according to the preset image feature set, and fill them into a vector that is initially empty, so as to obtain a sample image frame feature vector set; Obtain a feature extractor, where the feature extractor is constructed according to the mapping relationship between the sample image frame set and the sample image frame feature vector set; Use the feature extractor to extract features from the image frame sequence, and obtain the image frame feature vector sequence.

[0057] Furthermore, the target associated memory unit obtaining module 12 is used to perform the following steps: Extract a first image frame feature vector and a second image frame feature vector from the image frame feature vector sequence; Perform cross-frame feature correlation analysis on the first image frame feature vector and the second image frame feature vector to obtain a second correlated image frame feature vector, and add the first image frame feature vector and the second correlated image frame feature vector into the first associative memory unit; Extract a third image frame feature vector from the image frame feature vector sequence again, and use the first associative memory unit to perform cross-frame feature correlation memory update on the third image frame feature vector to obtain a second associative memory unit; Based on the second associative memory unit, perform cross-frame feature correlation analysis on the remaining image frame feature vectors of the image frame feature vector sequence in turn, and update the second associative memory unit according to the analysis results to obtain a target associative memory unit.

[0058] Further, the target associative memory unit obtaining module 12 is used to execute the following steps: Calculate the similarity of the corresponding elements of the first image frame feature vector and the second image frame feature vector to obtain an element similarity set; Perform normalization processing on the element similarity set, and matrixize the processed result to obtain an association analysis matrix; Use the association analysis matrix to enhance the second image frame feature vector to obtain a second correlated image frame feature vector.

[0059] Further, the target associative memory unit obtaining module 12 is used to execute the following steps: Use the association analysis matrix to perform weighted fusion on the second image frame feature vector to obtain a second enhanced image frame feature vector; Use the weighted average method to fuse the second enhanced image frame feature vector and the second image frame feature vector to obtain the second correlated image frame feature vector.

[0060] Further, the target associative memory unit obtaining module 12 is used to execute the following steps: Use the first image frame feature vector in the first associative memory unit to perform cross-frame feature correlation analysis on the third image frame feature vector to obtain a first initial third correlated image frame feature vector; Use the second correlated image frame feature vector in the first associative memory unit to perform cross-frame feature correlation analysis on the third image frame feature vector to obtain a second initial third correlated image frame feature vector; Perform mean processing on the first initial third correlated image frame feature vector and the second initial third correlated image frame feature vector to obtain a third correlated image frame feature vector, and add the third correlated image frame feature vector into the first associative memory unit to obtain a second associative memory unit.

[0061] Further, verify the target auxiliary review information, and optimize the network parameters of the target user queuing level identifier according to the verification result.

[0062] It should be noted that the above order of the embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above specific embodiments of this specification have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0063] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0064] This specification and the drawings are only exemplary descriptions of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A medical auxiliary review method for image recognition, characterized in that, The method includes: Interact with a medical imaging device connected to the hospital information system through an interface to export an image frame sequence of a target user within a preset window; Traverse the image frame sequence for cross-frame feature associated memory recognition update to obtain a target associated memory unit; Call a rank recognition device of the target user to extract information from the target associated memory unit, determine the rank of the target user, and use the rank of the target user as target auxiliary review information; Interact with the queuing management module of the hospital information system, perform department matching based on the target auxiliary review information, and obtain a target matching department; Perform a search for a queuing list based on the target matching department, combine the target auxiliary review information, and add the queuing application of the target user to the queuing list for optimization to obtain a target queuing list.

2. The medical auxiliary review method for image recognition according to claim 1, characterized in that Performing a search for a queuing list based on the target matching department, combining the target auxiliary review information, and adding the queuing application of the target user to the queuing list for optimization to obtain a target queuing list, includes: Traverse the queuing list to extract queuing auxiliary review information, and obtain a queuing auxiliary review information queue; Optimize the target auxiliary review information and the queuing auxiliary review information queue in descending order of the queuing rank to obtain the target queuing list.

3. The medical auxiliary review method for image recognition according to claim 1, wherein Traversing the image frame sequence for cross-frame feature associated memory recognition update to obtain a target associated memory unit, includes: Extract features from the image frame sequence respectively according to a preset image feature set to obtain an image frame feature vector sequence, where the preset image feature set includes structural texture features, edge information, shape features, and gray-scale density features; Perform cross-frame feature associated memory update based on the image frame feature vector sequence to obtain a target associated memory unit.

4. The medical assistance review method for image recognition according to claim 3, wherein Extract features from the image frame sequence respectively according to a preset image feature set, includes: Obtain a sample image frame set, extract the structural texture features, edge information, shape features, and gray-scale density features of each sample image frame in the sample image frame set respectively according to the preset image feature set, and fill them into an initially empty vector to obtain a sample image frame feature vector set; Obtain a feature extractor, where the feature extractor is constructed according to the mapping relationship between the sample image frame set and the sample image frame feature vector set; Use the feature extractor to extract features from the image frame sequence to obtain the image frame feature vector sequence.

5. The medical assistance review method for image recognition according to claim 3, wherein Perform cross-frame feature associated memory update based on the image frame feature vector sequence to obtain a target associated memory unit, includes: Extract a first image frame feature vector and a second image frame feature vector from the image frame feature vector sequence; Perform cross-frame feature associated analysis on the first image frame feature vector and the second image frame feature vector to obtain a second associated image frame feature vector, and add the first image frame feature vector and the second associated image frame feature vector to a first associated memory unit; Extract the third image frame feature vector from the image frame feature vector sequence again, and use the first associative memory unit to perform cross-frame feature associative memory update on the third image frame feature vector to obtain a second associative memory unit; Based on the second associative memory unit, perform cross-frame feature associative analysis on the remaining image frame feature vectors of the image frame feature vector sequence in turn, and update the second associative memory unit according to the analysis results to obtain a target associative memory unit.

6. The medical auxiliary review method for image recognition according to claim 5, wherein, Performing cross-frame feature associative analysis on the first image frame feature vector and the second image frame feature vector to obtain a second associated image frame feature vector, including: Calculate the similarity of the corresponding elements of the first image frame feature vector and the second image frame feature vector to obtain an element similarity set; Perform normalization processing on the element similarity set, and matrixize the processed result to obtain an association analysis matrix; Use the association analysis matrix to enhance the second image frame feature vector to obtain a second associated image frame feature vector.

7. The medical auxiliary review method for image recognition according to claim 6, characterized in that Including: Use the association analysis matrix to perform weighted fusion on the second image frame feature vector to obtain a second enhanced image frame feature vector; Use the weighted average method to fuse the second enhanced image frame feature vector and the second image frame feature vector to obtain the second associated image frame feature vector.

8. The medical auxiliary review method for image recognition according to claim 7, characterized in that, Extract the third image frame feature vector from the image frame feature vector sequence again, and use the first associative memory unit to perform cross-frame feature associative memory update on the third image frame feature vector to obtain a second associative memory unit, including: Use the first image frame feature vector in the first associative memory unit to perform cross-frame feature associative analysis on the third image frame feature vector to obtain a first initial third associated image frame feature vector; Use the second associated image frame feature vector in the first associative memory unit to perform cross-frame feature associative analysis on the third image frame feature vector to obtain a second initial third associated image frame feature vector; Perform mean processing on the first initial third associated image frame feature vector and the second initial third associated image frame feature vector to obtain a third associated image frame feature vector, and add the third associated image frame feature vector into the first associative memory unit to obtain a second associative memory unit.

9. The medical auxiliary review method for image recognition according to claim 1, characterized in that, Verify the target auxiliary review information, and optimize the network parameters of the target user queuing level recognizer according to the verification result.

10. A medical auxiliary review system for image recognition, characterized in that, The system is used to execute a medical auxiliary review method for image recognition according to any one of claims 1-9. The system includes: An image frame sequence export module, configured to interact with a medical imaging device connected to a hospital information system through an interface, and export an image frame sequence of a target user within a preset window; A target associative memory unit acquisition module, configured to traverse the image frame sequence for cross-frame feature associative memory recognition update to obtain a target associative memory unit; A target auxiliary review information acquisition module, configured to call a target user queuing level recognizer to extract information from the target associative memory unit, determine the target user queuing level, and use the target user queuing level as target auxiliary review information; A target matching department obtaining module, which is used to interact with the queuing management module of the hospital information system, perform department matching based on the target auxiliary review information, and obtain a target matching department; A target queuing queue obtaining module, which is used to perform a search for a queuing queue to be processed based on the target matching department, combine the target auxiliary review information, add the queuing application of the target user into the queuing queue to be processed for optimization, and obtain a target queuing queue.

Images