CN-122023073-A - Intelligent analysis teaching system for knee osteoarthritis images based on computer vision

CN122023073ACN 122023073 ACN122023073 ACN 122023073ACN-122023073-A

Abstract

The invention relates to the technical field of computer-aided teaching and medical image recognition, in particular to an intelligent analysis teaching system for knee osteoarthritis images based on computer vision, the system comprises an image data management module, a student interaction and behavior acquisition module, a multidimensional skill quantitative analysis module, a personalized knowledge graph construction and gap analysis module, a self-adaptive teaching path planning module and a teaching feedback and report generation module. By collecting and quantitatively analyzing the operation behaviors and the sight trajectories of students, constructing a personal knowledge graph to accurately position the skill defects and planning the personalized teaching path according to the personal knowledge graph, the fine assessment of the image analysis practice skills and the self-adaptive teaching guidance are realized.

Inventors

DING FENFANG
WU SI
SUN YIDAN
WANG JIE
NIE SHUANG
CHENG XIAOYU
LIU AIFENG
ZHANG CHAO
WU WEIYONG
XIE HAIBO

Assignees

天津中医药大学第一附属医院

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (10)

1. The utility model provides a knee joint osteoarthritis image intelligent analysis teaching system based on computer vision which characterized in that includes: the image data management module is used for storing and managing a teaching image database containing knee joint images, image marking information, standard diagnosis reports and image metadata; The student interaction and behavior acquisition module is used for presenting knee joint images to be analyzed in the graphical user interface and acquiring operation behavior sequences, sight focal tracks and time stamp information of students on the interaction interface in real time; The multidimensional skill quantitative analysis module is used for receiving the original behavior data acquired by the student interaction and behavior acquisition module, carrying out multidimensional quantitative evaluation on the image analysis skills of the students according to a preset quantitative model, and generating a skill quantitative report comprising an operation normalization score, an observation integrity score and a diagnosis logic consistency score; the personalized knowledge graph construction and gap analysis module is used for dynamically constructing a student personal knowledge graph representing the individual cognitive state of the student based on the skill quantitative report and standard knowledge in the teaching image database, and solving the knowledge weak points and skill defect items of the student by comparing the student personal knowledge graph with the standard expert knowledge graph; The self-adaptive teaching path planning module is used for receiving the difference analysis result output by the personalized knowledge graph construction and difference analysis module and generating a personalized follow-up learning content sequence and a training task instruction for the student based on a preset teaching resource library and a teaching strategy rule library; And the teaching feedback and report generation module is used for integrating the skill quantitative report and the personalized teaching path planning instruction, generating a visual learning report and driving a graphical user interface to carry out real-time feedback.
2. The intelligent analysis teaching system for knee osteoarthritis based on computer vision according to claim 1, wherein the multidimensional skill quantitative analysis module comprises an operation normalization analysis unit, an observation integrity analysis unit and a diagnostic logic consistency analysis unit; The operation normalization analysis unit is used for analyzing an operation behavior sequence, and the sequence is composed of a series of discrete interaction actions; the operation normalization analysis unit compares each interaction action with a normalization action template in a preset standard operation flow library, calculates action matching degree and records action sequence deviation degree; The observation integrity analysis unit is used for analyzing a sight line focus track, wherein the track consists of a series of coordinate points under an image coordinate system and corresponding time stamps, performing space-time correlation analysis on the sight line focus track and key pathological area marking information of the image in a teaching image database, calculating the total stay time ratio of the sight line in the key pathological area, and identifying the key area which is not covered by the sight line, wherein the observation integrity score is determined by the total stay time ratio and the number of the uncovered key areas; The diagnosis logic consistency analysis unit is used for carrying out association analysis on an operation behavior sequence, a sight focus track and a diagnosis conclusion text finally submitted by a student, firstly extracting an image feature set actually focused by the student from the behavior sequence and the sight track, analyzing an inference feature set relied by the student from the diagnosis conclusion text through a natural language processing technology, then calculating the intersection ratio of the actual focus feature set and the inference feature set, and analyzing whether an inference chain from the focus feature to the diagnosis conclusion accords with a preset medical diagnosis logic rule base or not, wherein the diagnosis logic consistency score is determined by the intersection ratio of the feature set and the degree that the inference chain accords with a logic rule together.
3. The intelligent analysis teaching system for knee osteoarthritis image based on computer vision according to claim 1, wherein the personalized knowledge graph construction and gap analysis module performs the process that the module maps each score in the skill quantification report to a mastery degree value of a corresponding node in the knowledge graph, the node represents a specific pathological sign recognition skill or diagnosis reasoning rule, meanwhile, the module updates the mastery degree value of the node by adopting an exponential weighted moving average algorithm according to the performance history data of the same type of node in multiple training of students so as to dynamically construct and update the personal knowledge graph of the students, the standard expert knowledge graph is predefined by field experts and comprises all knowledge nodes required by knee arthritis image analysis and logic relations among the nodes, the gap analysis is realized by comparing the mastery degree value of the corresponding node in the personal knowledge graph of the students with the mastery degree value of the standard expert knowledge graph node by node, and when the mastery degree value of a certain node is lower than a preset mastery degree threshold, the node is marked as a weak point, and meanwhile, the logic path of the associated weak node is analyzed, so that the core weak skill item is recognized.
4. The intelligent analysis teaching system for knee osteoarthritis based on computer vision according to claim 1, wherein the adaptive teaching path planning module comprises a teaching resource indexing unit and a policy rules engine; The teaching resource index unit is used for managing a structured teaching resource library, and each resource in the library is marked with a knowledge node label, a resource type and a difficulty level for training; The strategy rule engine is used for loading a preset teaching strategy rule base, the rule base comprises a series of production type rules, each rule is composed of preconditions and execution actions, the preconditions are set on the basis of types, severity and historical learning records of knowledge weak points and skill defect items in difference analysis results, and the execution actions are used for retrieving and combining teaching resources with specific labels, types and difficulty levels from a teaching resource index unit to form an ordered learning content sequence and training tasks.
5. The intelligent analysis teaching system for knee osteoarthritis based on computer vision according to claim 4, wherein the priority of rules in the teaching policy rule base is dynamically adjusted according to the severity of defect items, a high priority rule is triggered for defect items corresponding to key knowledge nodes with mastery level values lower than a first threshold value, a training task instruction generated by the rule is forcedly inserted into the forefront end of a current learning process, a medium priority rule is triggered for defect items corresponding to non-key knowledge nodes with mastery level values between the first threshold value and a second threshold value, the training task generated by the rule is arranged as recommended content in subsequent learning, and a consolidation rule is triggered for knowledge nodes with mastery level values higher than the second threshold value, wherein the training task is repeatedly arranged at intervals to prevent forgetting skills.
6. The computer vision-based intelligent analysis and teaching system for knee osteoarthritis imaging according to claim 1, wherein the system operates in a closed-loop workflow comprising 4 core links of learning, assessment, feedback, planning; In a learning link, a student interaction and behavior acquisition module presents teaching images and acquires behavior data; in the evaluation link, a multidimensional skill quantitative analysis module processes the behavior data to generate a skill quantitative report; In the feedback link, the teaching feedback and report generating module presents key components and a visual analysis chart in the skill quantification report to students in real time; In the planning link, the personalized knowledge graph construction and gap analysis module and the self-adaptive teaching path planning module work cooperatively based on the evaluation result to generate and update a learning plan of the next stage, so that the learning link is driven to enter a new iteration.
7. The computer vision-based intelligent analysis teaching system for knee osteoarthritis images of claim 1, further comprising a teaching effect tracking and model optimization module; The system comprises a module, a module and a manager, wherein the module is used for collecting and storing anonymized skill quantification reports, personal knowledge graph evolution sequences and final teaching result data of all students for a long time, the module periodically performs statistical analysis on the collected data, calculates validity indexes of different teaching strategy rules for improving specific types of skill defect items, and when the validity index of a teaching strategy rule is continuously lower than a preset optimization threshold value, the system gives an alarm to the manager to prompt the manager to examine and optimize the rule.
8. The intelligent analysis teaching system for knee osteoarthritis based on computer vision according to claim 2, wherein the operation normalization analysis unit aligns the actual operation sequence of the student with a normalization action template in a standard operation flow library by using a dynamic time normalization algorithm, calculates a matching degree score of each action, and the operation normalization score is determined by a weighted average of the matching degree scores of all steps together with a result obtained by multiplying a normalized deviation of the order of actions by a penalty factor.
9. The intelligent analysis teaching system for knee osteoarthritis image based on computer vision according to claim 2, wherein the observation integrity analysis unit calculates the observation integrity score by calculating the ratio of the total observation time length of the line of sight in all critical pathological areas to the total duration of the analysis task, and counting the number of critical areas not covered by any gazing event, and the observation integrity score is determined by the ratio and the number of uncovered critical areas together.
10. The intelligent analysis teaching system for knee osteoarthritis based on computer vision according to claim 3, wherein the process of updating the grasping degree value of the node by using the exponentially weighted moving average algorithm is as follows, namely, obtaining the sub item score corresponding to the node in the present training, and performing iterative updating by combining the grasping degree value of the node at the previous moment with a preset learning rate to obtain the grasping degree value of the node at the current moment.

Description

Intelligent analysis teaching system for knee osteoarthritis images based on computer vision Technical Field The invention belongs to the technical field of computer-aided teaching and medical image recognition, and particularly relates to an intelligent analysis teaching system for knee osteoarthritis images based on computer vision. Background Artificial intelligence technology is increasingly used in the field of medical health, and has great potential in the aspects of intelligent analysis and auxiliary diagnosis of medical images. Computer vision serves as a core branch of artificial intelligence, key features in medical images are automatically extracted and analyzed through a deep learning model, and a technical basis is provided for improving diagnosis and treatment efficiency and accuracy. Clinical skill teaching and assessment based on medical imaging is an important direction of medical education. Traditional teaching modes rely on teachers to explain standard case images, students learn disease features by observing and memorizing, and the core aim is to help students grasp the ability to identify pathological symptoms from the images and make correct diagnoses. The standard image library is adopted for teaching, and examination is carried out through selecting questions or simple answering questions, so that the mode has obvious limitation that the evaluation mode is highly dependent on static and preset answer options, and dynamic operation skills, observation paths and diagnosis reasoning logic of students in the actual image interpretation process cannot be effectively captured and quantified, so that skill evaluation flows on the surface and is one-sided. Meanwhile, the unified teaching content and assessment standard neglect the differences of students in knowledge mastering degree and skill weak links, and personalized training of teaching according to the material is difficult to realize. In the disease teaching of knee arthritis and the like which rely on image detail interpretation, the problems are particularly remarkable, so that the teaching effect is difficult to accurately measure and effectively improve. Disclosure of Invention The invention aims to provide an intelligent analysis teaching system for knee osteoarthritis images based on computer vision, which aims to solve the problems that in the prior art, a teaching evaluation mode cannot quantify dynamic operation skills and diagnosis reasoning logic and personalized teaching guidance is difficult to realize. The technical scheme of the invention is that the method comprises the following steps: the image data management module is used for storing and managing a teaching image database containing knee joint images, image marking information, standard diagnosis reports and image metadata; The student interaction and behavior acquisition module is used for presenting knee joint images to be analyzed in the graphical user interface and acquiring operation behavior sequences, sight focal tracks and time stamp information of students on the interaction interface in real time; The multidimensional skill quantitative analysis module is used for receiving the original behavior data acquired by the student interaction and behavior acquisition module, carrying out multidimensional quantitative evaluation on the image analysis skills of the students according to a preset quantitative model, and generating a skill quantitative report comprising an operation normalization score, an observation integrity score and a diagnosis logic consistency score; the personalized knowledge graph construction and gap analysis module is used for dynamically constructing a student personal knowledge graph representing the individual cognitive state of the student based on the skill quantitative report and standard knowledge in the teaching image database, and solving the knowledge weak points and skill defect items of the student by comparing the student personal knowledge graph with the standard expert knowledge graph; The self-adaptive teaching path planning module is used for receiving the difference analysis result output by the personalized knowledge graph construction and difference analysis module and generating a personalized follow-up learning content sequence and a training task instruction for the student based on a preset teaching resource library and a teaching strategy rule library; And the teaching feedback and report generation module is used for integrating the skill quantitative report and the personalized teaching path planning instruction, generating a visual learning report and driving a graphical user interface to carry out real-time feedback. Further, the multidimensional skill quantitative analysis module comprises an operation normalization analysis unit, an observation integrity analysis unit and a diagnostic logic consistency analysis unit. The operation normalization analysis unit is used for analyzing an operation behavior sequ