Automatic gait assessment method and system for cerebral palsy patient

Abstract

The invention belongs to the technical field of medicine and artificial intelligence, and discloses an automatic gait evaluation method and system for a cerebral palsy patient, and the system automatically captures and analyzes the walking lower limb movement characteristics of the patient through a computer vision technology, and evaluates gait abnormity and possible reasons thereof. The system aims at providing a tool which is low in cost, high in precision and easy to operate, and helps a doctor to objectively and effectively evaluate the walking function of a cerebral palsy patient. Abnormality evaluation and diagnosis are carried out on the lower limb walking function of the cerebral palsy patient. Walking data of a patient is recorded through a camera of the smart phone, and the system automatically identifies and analyzes motion characteristics of the lower limbs of the patient, such as joint angles, motion trails and space-time parameters, and identifies typical abnormal gaits to put forward possible reasons for abnormity.

Description

Technical Field

[0001] The present invention belongs to the technical fields of medicine and artificial intelligence, and particularly relates to a gait automatic assessment method and system for cerebral palsy patients. Background Art

[0002] Cerebral Palsy is a common neurological motor disorder, usually manifested as motor dysfunction and abnormal postures, and may be accompanied by abnormal muscle tone and movement coordination problems. Among cerebral palsy patients, motor dysfunction is a relatively common symptom, seriously affecting the quality of life and daily activity ability of patients. Traditional cerebral palsy assessment methods mainly rely on clinical observation and improved scoring systems, such as the Gross Motor Function Classification System (GMFCS) and the Manual Ability Classification System (MACS). These scales score by observing the patient's motor ability, postural control, and daily activity performance. However, these methods have certain limitations, mainly reflected in the following aspects:

[0003] Strong subjectivity: Traditional cerebral palsy motor function assessment relies on doctors' subjective judgments. There may be significant differences in the assessment results of the same patient by different doctors, resulting in poor repeatability of the assessment results. In addition, the fineness of existing scoring systems is limited and cannot capture subtle changes in movement, which may lead to the omission of some early or mild motor dysfunctions.

[0004] Time-consuming and low efficiency: Traditional assessment methods usually require doctors to judge the patient's motor function through long-term observation and recording. This not only places high time requirements on doctors but also may impose a large burden on patients. In some cases, patients need to repeat the same movement tasks multiple times to provide sufficient assessment information, thus reducing clinical efficiency.

[0005] Insufficient quantification ability: Traditional assessment methods are mainly based on visual observation and it is difficult to accurately quantify the patient's movement characteristics. Especially in subtle movements (such as joint angle changes), it is difficult for doctors to capture specific movement trajectories or dynamic changes with the naked eye. This lack of quantification ability limits the accuracy of the assessment and also hinders the formulation of personalized diagnosis and treatment plans.

[0006] Strong equipment dependence: Some improved assessment technologies such as three-dimensional motion capture systems, although able to provide a certain degree of quantitative support, usually require complex equipment and special environments. This not only increases the economic cost of the assessment but also limits its practical application in clinical rehabilitation and community rehabilitation settings.

[0007] To address the above problems, in recent years, automated motion assessment technologies based on computer vision have developed rapidly. Such technologies record the patient's motion videos through a camera and use pose estimation algorithms to extract the key points of human motion from them, and further analyze the motion characteristics. The advantage of this method lies in its automation, strong objectivity, and ability to efficiently quantify motion characteristics, providing more detailed motion data than traditional assessment methods.

[0008] Pose estimation algorithms in computer vision technologies, such as MediaPipe and OpenPose, can detect multiple key points of the human body (such as the hip, knee, ankle, etc.) from a video through a single camera. These key points can reflect the postural changes and joint range of motion of the patient during exercise. Compared with traditional motion capture systems, this vision-based markerless technology has the advantages of low cost, strong portability, and wide applicability, and is very suitable for use in clinical environments.

[0009] In addition, current clinical therapists generally lack biomechanical knowledge related to gait analysis, and the abnormal gait of cerebral palsy patients often may involve complex mechanical, neurological, and compensatory factors. Accurately identifying these abnormalities and their potential causes is challenging for therapists. Therefore, the current medical environment requires systematic gait analysis and diagnostic tools, which can not only significantly reduce the burden on therapists, but also provide an objective basis for treatment strategies, further improving the rehabilitation effect and clinical efficiency.

[0010] Through the above analysis, the problems and defects of the existing technology are as follows:

[0011] (1) Strong subjectivity: Traditional assessment of cerebral palsy motor function relies on the subjective judgment of doctors. There may be significant differences in the assessment results of the same patient by different doctors, resulting in poor repeatability of the assessment results. In addition, the fineness of the existing scoring system is limited and cannot capture the subtle changes in motion, which may lead to the omission of some early or mild motor function disorders.

[0012] (2) Time-consuming and low efficiency: Traditional assessment methods usually require doctors to judge the patient's motor function through long-term observation and recording. This not only places high time requirements on doctors, but also may impose a greater burden on patients. In some cases, patients need to repeat the same motion tasks multiple times to provide sufficient assessment information, thus reducing clinical efficiency.

[0013] (3) Insufficient quantification ability: Traditional assessment methods are mainly based on visual observation and it is difficult to accurately quantify the patient's motion characteristics. Especially in subtle motions (such as joint angle changes), it is difficult for doctors to capture the specific motion trajectory or dynamic changes with the naked eye. This lack of quantification ability limits the accuracy of the assessment and also hinders the formulation of personalized diagnosis and treatment plans.

[0014] (4) Strong device dependence: Some improved evaluation technologies, such as three-dimensional motion capture systems, although they can provide a certain degree of quantitative support, usually require complex devices and special environments. This not only increases the economic cost of evaluation but also limits its practical application in clinical rehabilitation and community rehabilitation settings.

[0015] (5) Insufficient ability to specifically explain cerebral palsy: Existing methods lack the ability to provide targeted explanations for gait abnormalities in cerebral palsy, including mechanical, neurological, and compensatory factors. This problem of insufficient explanatory ability limits the accuracy of evaluation and clinical evaluation efficiency. Summary of the Invention

[0016] In view of the problems existing in the prior art, the present invention provides a method and system for automatic gait evaluation of cerebral palsy patients.

[0017] The present invention is implemented as follows. An automatic gait evaluation method for cerebral palsy patients includes:

[0018] Step 1, data collection;

[0019] In the first step of system operation, the data collection module is responsible for recording the patient's walking process through the smartphone camera; the patient needs to walk on a prescribed path; the walking task can be carried out in a hospital, rehabilitation center, or the patient's daily living environment;

[0020] Step 2, pose estimation and key point detection;

[0021] After the data collection is completed, the system enters the pose estimation and key point detection module; uses the open-source MediaPipe computer vision algorithm to automatically detect and track the movement of the patient's lower limbs from the video; identifies 33 key points of the human body through a convolutional neural network (CNN), including the hip, knee, ankle, foot, and parts related to lower limb movement;

[0022] Step 3, gait cycle division;

[0023] After the pose estimation is completed and the key point data is obtained, the system enters the gait cycle division module; extracts the complete gait cycle according to the identified gait events for subsequent analysis;

[0024] Step 4, gait parameter analysis;

[0025] After identifying the gait events and obtaining the gait cycle, the system enters the gait parameter analysis module; calculates gait spatio-temporal parameters and kinematic parameters; these characteristic parameters provide a reference for therapists to evaluate the patient's walking performance;

[0026] Step 5, gait abnormality diagnosis;

[0027] After obtaining the gait parameters, the system enters the gait abnormality diagnosis module; automatically identifies gait abnormalities and infers their possible causes; this information provides a reference for therapists to analyze the gait parameter results and make a diagnosis.

[0028] Step 6, result output and report generation.

[0029] After completing the abnormal gait diagnosis, the system enters the report generation module, providing detailed and storable evaluation results for clinicians; the evaluation report includes the joint angle change curve during walking, spatio-temporal parameter results, and the abnormal diagnosis and possible causes of joint angles.

[0030] Step 7, cloud deployment and system maintenance.

[0031] Furthermore, the data collection:

[0032] Device confirmation:

[0033] Smartphone, a 4-meter-long walking path space, a tripod or a phone holder for fixing the phone.

[0034] Device placement location:

[0035] 1) The phone is fixed on a tripod or a phone holder, and the lens height is at the hip level of the child.

[0036] 2) The phone is on one side of the walking path, placed horizontally and perpendicular to the ground, with the rear camera facing the walking path, taking a side view of the child walking. The phone is about 3.5 meters away from the center of the walking path.

[0037] Furthermore, the gait test:

[0038] Before the test, the participant first walks habitually on the walking path to adapt to the environment; during the formal test, the participant takes off their shoes and walks along a straight track at a normal speed, collecting videos of walking from left to right or from right to left; ensure that the entire body of the participant is captured during the video shooting, and the video needs to collect data of at least 3 gait cycles.

[0039] Furthermore, the pose estimation and key point detection:

[0040] 1) Pose estimation algorithm: The MediaPipe model automatically detects key points in the whole body movement of the patient based on the RGB images captured by the camera; these key points are represented by three-dimensional coordinates x, y, z. The x and y coordinates represent the position of the key points in the plane, and the z coordinate is used to estimate the depth of the key points; through the pose estimation algorithm, the system can real-time track the lower limb movement trajectory of the patient.

[0041] 2) Data preprocessing: After obtaining the key point coordinate data, the system preprocesses the original data; this step includes removing noise and filtering abnormal data points; to avoid interference from camera jitter or light changes on the detection results, the system smooths the coordinate changes between frames.

[0042] Furthermore, the gait parameter analysis:

[0043] 1) Spatiotemporal parameter analysis: The system calculates gait spatiotemporal parameters based on the position information obtained from posture assessment, such as: walking speed, step frequency, step length;

[0044] 2) Joint angle analysis: The system tracks and analyzes the joint angles of the hip, knee, and ankle; by calculating the angle changes of the joints within the sub-stages of the gait cycle;

[0045] Furthermore, the gait abnormality diagnosis:

[0046] Abnormality recognition: The system compares the change curves of joint angles in the sub-stages of the gait cycle with the normal population dataset to identify typical abnormalities of cerebral palsy under single joints; then, based on the built-in knowledge base, the system infers the mechanical, nervous system, and musculoskeletal causes that may trigger these abnormalities.

[0047] Furthermore, the result output and report generation:

[0048] Report content: The report details the spatiotemporal parameter information of the patient in tabular form, facilitating the therapist to comprehensively evaluate the walking function; the evaluation results of joint angles are presented in a curve graph, intuitively showing the patient's movement characteristics, and accompanied by a comparison with the normal gait curve to help quickly identify potential abnormalities; the gait abnormalities detected by the automatic algorithm and their possible causes are supplemented with clear explanations, providing an objective reference for the therapist and assisting in formulating more accurate rehabilitation strategies;

[0049] Furthermore, the cloud deployment and system maintenance:

[0050] Cloud deployment: The computing and storage resources of the system are hosted in the cloud, capable of processing the data of multiple patients simultaneously; users only need to upload the video, and the cloud system will automatically analyze and return the evaluation report, which is simple to operate and suitable for rapid application in the clinical environment;

[0051] System maintenance and update: The system supports post-maintenance and update, and can be continuously updated with the development of the posture recognition algorithm.

[0052] The purpose of the present invention is to provide an automated gait assessment system for cerebral palsy patients, including:

[0053] The data acquisition module is used to, in the first step of system operation, be responsible for recording the patient's walking process through the smartphone camera; the patient needs to walk on a prescribed walking path; the walking task can be carried out in a hospital, a rehabilitation center, or the patient's daily living environment.

[0054] The detection module is used for pose estimation and key point detection; after the data acquisition is completed, the system enters the pose estimation and key point detection module; using the open-source MediaPipe computer vision algorithm, it automatically detects and tracks the movement of the patient's lower limbs from the video; 33 key points of the human body are identified through a convolutional neural network (CNN), including the hip, knee, ankle, foot, and parts related to the movement of the lower limbs.

[0055] The division module is used for gait cycle division; after the pose estimation is completed and the key point data is obtained, the system enters the gait cycle division module; complete gait cycles are extracted according to the identified gait events for subsequent analysis.

[0056] The analysis module is used for gait parameter analysis; after identifying the gait events and obtaining the gait cycles, the system enters the gait parameter analysis module; gait spatio-temporal parameters and kinematic parameters are calculated; these characteristic parameters provide a reference for therapists to evaluate the patient's walking performance.

[0057] The diagnosis module is used for gait abnormality diagnosis; after obtaining the gait parameters, the system enters the gait abnormality diagnosis module; gait abnormalities are automatically identified and their possible causes are recommended; this information provides a reference for therapists to analyze the gait parameter results and make a diagnosis.

[0058] The output module is used for result output and report generation; after completing the abnormal gait diagnosis, the system enters the report generation module to provide detailed evaluation results for clinicians; the evaluation report includes the joint angle change curve during the walking process, the spatio-temporal parameter results, and the abnormal diagnosis and possible causes of the joint angles.

[0059] The maintenance module is used for cloud deployment and system maintenance.

[0060] Another object of the present invention is to provide a computer device, the computer device includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for automatic evaluation of the gait of cerebral palsy patients.

[0061] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the processor executes the steps of the method for automatic evaluation of the gait of cerebral palsy patients.

[0062] Another object of the present invention is to provide an information data processing terminal for implementing the gait automatic evaluation system for cerebral palsy patients.

[0063] Combined with the above technical solutions and the solved technical problems, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0064] First, the present invention proposes an automatic cerebral palsy gait evaluation system based on computer vision technology, which is particularly used for abnormally evaluating and diagnosing the lower limb walking function of cerebral palsy patients. By recording the walking data of patients through the smartphone camera, the system automatically identifies and analyzes the movement characteristics of the patients' lower limbs, such as joint angles, movement trajectories, spatio-temporal parameters, etc., and identifies typical abnormal gaits and proposes possible causes of the abnormalities. By inputting these into the patients' gait videos, the system can automatically perform gait analysis to generate a gait analysis report, helping doctors accurately evaluate the patients' motor functions and providing references for the possible causes of the abnormalities.

[0065] Compared with the traditional evaluation methods, the present invention has the following remarkable advantages:

[0066] 1. Objectivity and accuracy: The system evaluates in a data-driven manner, reducing the deviation caused by subjective judgment, being able to capture subtle changes during movement, and providing accurate quantitative data.

[0067] 2. High-efficiency automation: The system can quickly process a large amount of movement data, reducing the workload of doctors and significantly improving the efficiency of clinical evaluation.

[0068] 3. Convenience and low cost: This system does not require special hardware devices. Only by recording the videos of patients through the smartphone camera can the evaluation be completed, which is suitable for a wide range of clinical and rehabilitation environments.

[0069] In summary, the technology of the present invention can effectively solve the limitations in the prior art, is particularly suitable for the gait evaluation of cerebral palsy patients, and provides an objective, convenient and efficient evaluation tool for clinicians.

[0070] The present invention provides a method and system for automatic gait evaluation of cerebral palsy patients. The system automatically captures and analyzes the movement characteristics of the lower limbs of patients walking by using computer vision technology, evaluates gait abnormalities and their possible causes. The system aims to provide a low-cost, high-precision and easy-to-operate tool to help doctors objectively and effectively evaluate the walking function of cerebral palsy patients.

[0071] Second, as the creative auxiliary evidence of the claims of the present invention, it is also reflected in the following important aspects:

[0072] (1) The expected benefits and commercial values after the transformation of the technical solutions of the present invention are:

[0073] By providing an intelligent gait assessment solution, the dependence on expensive hardware and professional equipment in traditional assessment methods is significantly reduced, thus remarkably lowering the implementation cost and having the potential for large-scale clinical promotion and application. The convenience and efficiency of the system contribute to its application in multiple scenarios such as rehabilitation centers and home environments, promoting the development of remote rehabilitation and follow-up services.

[0074] (2) The technical solution of the present invention overcomes the technical prejudice: Traditional gait assessment methods are often limited to the laboratory environment and rely on professional equipment and complex processes, resulting in their inability to be popularized in clinical scenarios with limited resources. The present invention breaks through this technical prejudice. Through software algorithm optimization, ordinary smartphone devices can achieve accurate and rapid assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 is a flowchart of the automatic gait assessment method for cerebral palsy patients provided by an embodiment of the present invention.

[0076] Figure 2 is a block diagram of the automatic gait assessment system for cerebral palsy patients provided by an embodiment of the present invention.

[0077] Figure 3 is a schematic diagram of device connection provided by an embodiment of the present invention.

[0078] Figure 4 is a schematic diagram of the implementation process provided by an embodiment of the present invention.

[0079] Figure 5 is a diagram of the data display page on the software side provided by an embodiment of the present invention.

[0080] Figure 6 is a data display page - posture recognition diagram provided by an embodiment of the present invention.

[0081] Figure 7 is a data display page - gait cycle recognition diagram provided by an embodiment of the present invention.

[0082] Figure 8 is a data display page - kinematic analysis diagram provided by an embodiment of the present invention.

[0083] Figure 9 is a data display page - gait abnormality diagnosis diagram provided by an embodiment of the present invention.

[0084] Figure 10 is a diagram of the accuracy of gait abnormality diagnosis provided by an embodiment of the present invention.

[0085] Figure 11 is a 5-component scale diagram of the system feasibility assessment provided by an embodiment of the present invention. Detailed implementation mode

[0086] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the following further describes the present invention in detail with reference to embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0087] As Figure 1 shown, a method for automatic gait assessment for cerebral palsy patients provided by an embodiment of the present invention includes the following steps:

[0088] S101, data collection;

[0089] In the first step of system operation, the data collection module is responsible for recording the patient's walking process through a smartphone camera; the patient needs to walk on a specified walking path; the walking task can be carried out in a hospital, a rehabilitation center or the patient's daily living environment;

[0090] S102, pose estimation and key point detection;

[0091] After the data collection is completed, the system enters the pose estimation and key point detection module; uses the open-source MediaPipe computer vision algorithm to automatically detect and track the movement of the patient's lower limbs from the video; identifies 33 key points of the human body through a convolutional neural network (CNN), including the hip, knee, ankle, foot and other parts related to the movement of the lower limbs;

[0092] S103, gait cycle division;

[0093] After the pose estimation is completed and the key point data is obtained, the system enters the gait cycle division module; extracts the complete gait cycle according to the identified gait events for subsequent analysis;

[0094] S104, gait parameter analysis;

[0095] After identifying the gait events and obtaining the gait cycle, the system enters the gait parameter analysis module; calculates the spatio-temporal parameters and kinematic parameters of the gait; these characteristic parameters provide a reference for the therapist to evaluate the patient's walking performance;

[0096] S105, gait abnormality diagnosis;

[0097] After obtaining the gait parameters, the system enters the gait abnormality diagnosis module; automatically identifies gait abnormalities and recommends their possible causes; this information provides a reference for the therapist to analyze the gait parameter results and make a diagnosis;

[0098] S106, result output and report generation;

[0099] After completing the abnormal gait diagnosis, the system enters the report generation module to provide detailed evaluation results for clinicians; the evaluation report includes the joint angle change curve during walking, spatio-temporal parameter results, as well as the abnormal diagnosis and possible causes of joint angles;

[0100] S107, Cloud deployment and system maintenance.

[0101] Data acquisition provided by the embodiments of the present invention:

[0102] Device confirmation:

[0103] A smartphone, a 4-meter-long walking path space, and a tripod or phone holder for fixing the phone;

[0104] Device placement location:

[0105] 1) The phone is fixed on a tripod or phone holder, and the lens height is at the hip level of the child;

[0106] 2) The phone is located on one side of the walking path, placed horizontally and perpendicular to the ground, with the rear camera facing the walking path, and the phone is about 3.5 meters away from the center of the walking path to capture the side view of the child walking.

[0107] Gait test provided by the embodiments of the present invention:

[0108] Before the test, the participant first walks habitually on the walking path to adapt to the environment; during the formal test, the participant takes off their shoes and walks at a normal speed along a straight track, collecting videos of walking from left to right or from right to left; ensure that the entire body of the participant is captured during the video shooting, and the video needs to collect data of at least 3 gait cycles.

[0109] Pose estimation and key point detection provided by the embodiments of the present invention:

[0110] 1) Pose estimation algorithm: The MediaPipe model automatically detects key points in the full-body movement of the patient based on the RGB images captured by the camera; these key points are represented by three-dimensional coordinates of x, y, and z. The x and y coordinates represent the position of the key point in the plane, and the z coordinate is used to estimate the depth of the key point; through the pose estimation algorithm, the system can real-time track the lower limb movement trajectory of the patient;

[0111] 2) Data preprocessing: After obtaining the key point coordinate data, the system preprocesses the original data; this step includes removing noise and filtering abnormal data points; to avoid interference from camera jitter or light changes on the detection results, the system smooths the coordinate changes between frames.

[0112] Gait parameter analysis provided by the embodiments of the present invention:

[0113] 1) Spatiotemporal parameter analysis: The system calculates gait spatiotemporal parameters such as walking speed, cadence, and step length based on the position information obtained from the posture assessment.

[0114] 2) Joint angle analysis: The system tracks and analyzes the joint angles of the hip, knee, and ankle; by calculating the angle changes of the joints within the sub - phases of the gait cycle.

[0115] Gait abnormality diagnosis provided by the embodiments of the present invention:

[0116] Abnormality recognition: The system compares the change curves of joint angles in the sub - phases of the gait cycle with the normal population dataset to identify typical abnormalities of cerebral palsy under single joints; then, based on the built - in knowledge base, the system infers the mechanical, nervous system, and musculoskeletal causes that may trigger these abnormalities.

[0117] Result output and report generation provided by the embodiments of the present invention:

[0118] Report content: The report details the spatiotemporal parameter information of the patient in tabular form, facilitating the therapist to comprehensively evaluate the walking function; the evaluation results of joint angles are presented in a curve graph, intuitively showing the patient's movement characteristics, and accompanied by a comparison with the normal gait curve to help quickly identify potential abnormalities; the gait abnormalities detected by the automatic algorithm and their possible causes are supplemented with clear explanations, providing an objective reference for the therapist and assisting in formulating more accurate rehabilitation strategies.

[0119] Cloud deployment and system maintenance provided by the embodiments of the present invention:

[0120] Cloud deployment: The computing and storage resources of the system are hosted in the cloud, capable of processing the data of multiple patients simultaneously; users only need to upload the video, and the cloud system will automatically analyze and return the evaluation report, with simple operation, suitable for rapid application in the clinical environment.

[0121] System maintenance and update: The system supports post - maintenance and update, and can be continuously updated with the development of the posture recognition algorithm.

[0122] Such as Figure 2 As shown, an automated gait assessment system for cerebral palsy patients provided by the embodiments of the present invention includes:

[0123] A data acquisition module, which is responsible for recording the patient's walking process through the smartphone camera in the first step of the system operation; the patient needs to walk on a prescribed path; the walking task can be carried out in a hospital, a rehabilitation center, or the patient's daily living environment.

[0124] The detection module is used for pose estimation and key point detection. After the data acquisition is completed, the system enters the pose estimation and key point detection module. Using the open-source MediaPipe computer vision algorithm, it automatically detects and tracks the movement of the patient's lower limbs from the video. It identifies 33 key points of the human body through a convolutional neural network (CNN), including the hip, knee, ankle, foot and parts related to the movement of the lower limbs.

[0125] The division module is used for gait cycle division. After the pose estimation is completed and the key point data is obtained, the system enters the gait cycle division module. It extracts the complete gait cycle according to the recognized gait events for subsequent analysis.

[0126] The analysis module is used for gait parameter analysis. After the gait events are recognized and the gait cycle is obtained, the system enters the gait parameter analysis module. It calculates the spatio-temporal parameters and kinematic parameters of the gait. These characteristic parameters provide a reference for therapists to evaluate the walking performance of patients.

[0127] The diagnosis module is used for gait abnormality diagnosis. After the gait parameters are obtained, the system enters the gait abnormality diagnosis module. It automatically identifies gait abnormalities and recommends their possible causes. These information provides a reference for therapists to analyze the gait parameter results and make a diagnosis.

[0128] The output module is used for result output and report generation. After the abnormal gait diagnosis is completed, the system enters the report generation module to provide detailed evaluation results for clinicians. The evaluation report includes the joint angle change curve during walking, the spatio-temporal parameter results, and the abnormal diagnosis and possible causes of the joint angles.

[0129] The maintenance module is used for cloud deployment and system maintenance.

[0130] Another object of the present invention is to provide a computer device, the computer device includes a memory and a processor, the memory stores a computer program, when the computer program is executed by the processor, the processor executes the steps of the gait automatic evaluation method for cerebral palsy patients.

[0131] Another object of the present invention is to provide a computer-readable storage medium, storing a computer program, when the computer program is executed by a processor, the processor executes the steps of the gait automatic evaluation method for cerebral palsy patients.

[0132] Another object of the present invention is to provide an information data processing terminal, the information data processing terminal is used to implement the gait automatic evaluation system for cerebral palsy patients.

[0133] The specific implementation of the present invention:

[0134] The present invention provides a method and system for automatic gait assessment of cerebral palsy patients. The system uses computer vision technology to automatically capture and analyze the lower limb movement characteristics of the patient's walking, and evaluate gait abnormalities and their possible causes. The system aims to provide a low-cost, high-precision, and easy-to-operate tool to help doctors objectively and effectively evaluate the walking function of cerebral palsy patients. The following is a detailed description of each step in the method of the present invention.

[0135] Data collection

[0136] As Figure 3 , Figure 4 shown, in the first step of the system operation, the data collection module is responsible for recording the patient's walking process through the smartphone camera. The patient needs to walk on a prescribed walking path. The walking task can be carried out in a hospital, a rehabilitation center, or the patient's daily living environment.

[0137] 1. Equipment confirmation

[0138] Smartphone (adjust the shooting video parameters to at least 30Hz 1080p), a 4-meter-long walking path space, a tripod or a mobile phone holder for fixing the mobile phone

[0139] 2. Equipment placement location

[0140] 1) Fix the mobile phone on the tripod or the mobile phone holder, and the lens height is at the hip level of the child (about 0.5 - 0.8 meters)

[0141] 2) The mobile phone is located on one side of the walking path, placed horizontally and perpendicular to the ground, and the rear camera faces the walking path, shooting the side view of the child walking. The mobile phone is about 3.5 meters away from the center of the walking path (it is necessary to ensure that the whole body of the participant and the full length of the walking path can be completely captured).

[0142] 3. Gait test

[0143] Before the test, the participant first walks habitually on the walking path to adapt to the environment. During the formal test, the participant takes off shoes and walks at a normal speed along a straight track, and collects videos of walking from left to right or from right to left. Ensure that the complete body of the participant is captured during the video shooting process, and the video needs to collect data of at least 3 gait cycles.

[0144] Pose estimation and key point detection

[0145] As Figure 5 , Figure 6 shows that after the data collection is completed, the system enters the pose estimation and key point detection module. The system uses open-source computer vision algorithms such as MediaPipe to automatically detect and track the movement of the patient's lower limbs from the video. These algorithms identify 33 key points of the human body through a convolutional neural network (CNN), including parts related to lower limb movement such as the hip, knee, ankle, and foot.

[0146] 1. Pose Estimation Algorithm: The MediaPipe model automatically detects key points in the patient's full-body movement based on the RGB images captured by the camera. These key points are represented by three-dimensional coordinates x, y, z, where the x and y coordinates represent the position of the key points in the plane, and the z coordinate is used to estimate the depth of the key points. Through the pose estimation algorithm, the system can track the lower limb movement trajectory of the patient in real time.

[0147] 2. Data Preprocessing: After obtaining the key point coordinate data, the system preprocesses the original data. This step includes removing noise and filtering out abnormal data points. To avoid interference from camera jitter or light changes on the detection results, the system smooths the coordinate changes between frames to ensure the continuity and accuracy of the data.

[0148] Gait Cycle Division

[0149] As Figure 7 shown, after the pose estimation is completed and the key point data is obtained, the system enters the gait cycle division module. The goal of this step is to extract the complete gait cycle based on the recognized gait events for subsequent analysis.

[0150] To identify important gait events such as foot strike and foot off, this system uses a customized algorithm that can detect the farthest forward and backward positions of the foot relative to the hip. Then the filtered position data is used to detect gait events such as foot strike and foot off. The system supports both automatic and manual segmentation. When the toe point reaches the farthest forward position (the relative distance between the hip and the toe point is the largest), the foot strike event is detected. When the toe point is in the rearmost position (the relative distance between the hip and the toe point is the smallest), the foot off event is determined. The system also supports manual segmentation, allowing users to visually check and record the frame numbers as needed. After identifying the foot strike, the gait cycle is segmented and linearly interpolated to 101 frames. The gait cycle is further divided into sub-phases according to the foot off moment: the first double support phase, the single support phase, the second double support phase, and the swing phase.

[0151] Gait Parameter Analysis

[0152] As Figure 8 shown, after identifying the gait events and obtaining the gait cycle, the system enters the gait parameter analysis module. The purpose of this step is to calculate the spatio-temporal parameters and kinematic parameters of gait. These characteristic parameters provide a reference for therapists to evaluate the patient's walking performance.

[0153] 1. Spatiotemporal parameter analysis: The system calculates gait spatiotemporal parameters, such as walking speed, cadence, step length, etc., based on the position information obtained from pose assessment. To avoid the influence of individual body size, camera position, and image depth on data calculation, the system is calibrated using height. Spatiotemporal parameters can reflect the overall characteristics and patterns of the patient's gait and are important indicators for evaluating gait function.

[0154] 2. Joint angle analysis: The system tracks and analyzes the joint angles of the hip, knee, and ankle. By calculating the angle changes of the joints within the sub - phases of the gait cycle, the system can detect whether there are any abnormal manifestations in the movement of each body segment. The changes in joint angles are an important indicator in gait assessment, especially when the patient performs repetitive movements, and the abnormal changes in angles can often clearly reflect problems.

[0155] Gait abnormality diagnosis

[0156] As Figure 9 shown, after obtaining the gait parameters, the system enters the gait abnormality diagnosis module. The purpose of this step is to automatically identify gait abnormalities and recommend their possible causes. This information provides a reference for therapists to analyze the gait parameter results and make diagnoses.

[0157] 1. Abnormality identification: The system compares the change curves of joint angles in the sub - phases of the gait cycle with the normal population dataset to identify typical abnormalities of cerebral palsy under single joints. This system has developed a rule - based automatic identification algorithm, which has identified 11 gait abnormalities, including "excessive knee flexion during the first double - support phase", around the sagittal plane of the lower - limb hip, knee, and ankle.

[0158] 2. Possible causes: Based on previously published studies by predecessors and focus group discussions with experienced clinical therapists, we have sorted out the possible causes of cerebral palsy gait abnormalities. After automatically identifying gait abnormalities, the system will call the corresponding possible causes as explanations for the abnormalities.

[0159] Result output and report generation

[0160] After completing the abnormal gait diagnosis, the system enters the report generation module to provide detailed assessment results for clinicians. The assessment report includes the change curves of joint angles during walking, spatiotemporal parameter results, as well as the abnormal diagnosis and possible causes of joint angles.

[0161] Report content: The report details the patient's spatiotemporal parameter information in tabular form, facilitating therapists to comprehensively evaluate walking function. The assessment results of joint angles are presented as curves, intuitively showing the patient's movement characteristics, and are accompanied by a comparison with the normal gait curve to help quickly identify potential abnormalities. The gait abnormalities detected by the automatic algorithm and their possible causes are supplemented with clear explanations, providing an objective reference for therapists and assisting in formulating more precise rehabilitation strategies.

[0162] Cloud Deployment and System Maintenance

[0163] The evaluation system of the present invention can be deployed in the cloud. Users only need to upload the video data of the patient through a smartphone or computer, and the system can complete data processing and analysis in the cloud. The system supports multiple users to use simultaneously and can be widely applied in environments such as hospitals and rehabilitation centers.

[0164] 1. Cloud Deployment: The computing and storage resources of the system are hosted in the cloud, which can process the data of multiple patients simultaneously. Users only need to upload the video, and the cloud system will automatically analyze and return the evaluation report. The operation is simple and suitable for rapid application in the clinical environment.

[0165] 2. System Maintenance and Update: The system supports later maintenance and update, and can be continuously updated with the development of the pose recognition algorithm to obtain more accurate and reliable evaluation results. The parameter adjustment and update of the pose recognition algorithm model can be carried out through training with the data of cerebral palsy patients to better identify the typical abnormal gaits of cerebral palsy patients.

[0166] I. The specific application fields or related products of the present invention.

[0167] Examples of actual application scenarios:

[0168] Clinical Scenario: The patient conducts a gait test in the evaluation room of the rehabilitation center, and uses a single camera to record the gait (either in real-time or by recording a video and uploading it to the cloud). The therapist can obtain a detailed data report through the evaluation system.

[0169] Home Scenario: The patient uses the camera of the mobile phone to record a video in the corridor or living room, uploads it to the cloud, and the system conducts automated analysis and provides feedback reports.

[0170] Remote Rehabilitation Scenario: The gait of the patient is live-streamed through a single camera, and gait parameters are generated in real-time and shared remotely with the doctor.

[0171] Examples of actual application devices:

[0172] In this embodiment, the gait evaluation system is deployed in a web-based application or a smartphone application. The gait video of the patient is collected through a camera, and a lightweight algorithm is used to analyze the video data in real-time. On devices with lower hardware performance, the algorithm adopts a simplified model to reduce the amount of calculation, achieve efficient operation, and ensure the evaluation accuracy.

[0173] II. Relevant evidence for obtaining technical effects in the embodiments of the present invention.

[0174] Using marker-based three-dimensional motion capture as the gold standard, the accuracy rate of gait abnormality recognition by the system of the present invention was compared with the diagnostic result accuracy rate of therapists' observational gait analysis. Gait data of 18 cerebral palsy patients were used for the comparison, and the results are shown in the bar Figure 10 , indicating that the accuracy of gait abnormality recognition by the system of the present invention is better than that of therapists' observational gait analysis.

[0175] Nineteen therapists specializing in the care of children with cerebral palsy completed a customized 5-point scale (1 = very dissatisfied, 5 = very satisfied) after using the system of the present invention, and evaluated the clinical application feasibility of the system from four aspects: whether the system interface design is reasonable, whether the system diagnosis result is consistent with their own observation, whether it is clinically practical, and whether there is an intention to use it in the future. The bar Figure 11 The results showed that the system scores were all higher than 4, between 4.42 and 4.89, indicating that the system has high clinical usability.

[0176] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. Those of ordinary skill in the art can understand that the above-mentioned devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code is provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and their modules of the present invention can be implemented by hardware circuits of programmable hardware devices such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips and transistors, or field programmable gate arrays and programmable logic devices, can also be implemented by software executed by various types of processors, or can be implemented by a combination of the above hardware circuits and software such as firmware.

[0177] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A gait automatic assessment method for cerebral palsy patients, characterized in that, It includes the following steps: Step 1, data collection; In the first step of system operation, the data collection module is responsible for recording the patient's walking process through the smartphone camera; The patient needs to walk on a specified walking path; the walking task can be carried out in a hospital, a rehabilitation center or the patient's daily living environment; Step 2, pose estimation and key point detection; After the data collection is completed, the system enters the pose estimation and key point detection module; using the open-source MediaPipe computer vision algorithm, it automatically detects and tracks the movement of the patient's lower limbs from the video; 33 key points of the human body are identified through a convolutional neural network (CNN), including the hip, knee, ankle, foot and other parts related to the movement of the lower limbs; Step 3, gait cycle division; After the pose estimation is completed and the key point data is obtained, the system enters the gait cycle division module; extracts the complete gait cycle according to the identified gait events for subsequent analysis; Step 4, gait parameter analysis; After identifying the gait events and obtaining the gait cycle, the system enters the gait parameter analysis module; Calculates gait spatio-temporal parameters and kinematic parameters; These characteristic parameters provide reference for therapists to evaluate the patient's walking performance; Step 5, gait abnormality diagnosis; After obtaining the gait parameters, the system enters the gait abnormality diagnosis module; automatically identifies gait abnormalities and recommends their possible causes; this information provides reference for therapists to analyze the gait parameter results and make a diagnosis; Step 6, result output and report generation; After completing the abnormal gait diagnosis, the system enters the report generation module, providing detailed evaluation results for clinicians; the evaluation report includes the joint angle change curve during walking, the spatio-temporal parameter results, and the abnormal diagnosis and possible causes of joint angles; Step 7, cloud deployment and system maintenance.

2. The gait automatic evaluation method for cerebral palsy patients according to claim 1, wherein The data collection: Device confirmation: Smartphone, a 4-meter-long walking path space, a tripod or a mobile phone holder for fixing the mobile phone; Device placement location: 1) The mobile phone is fixed on the tripod or mobile phone holder, and the lens height is at the hip level of the child; 2) The mobile phone is located on one side of the walking path, placed horizontally and perpendicular to the ground, with the rear camera facing the walking path, and the mobile phone is about 3.5 meters away from the center of the walking path, shooting the side view of the child walking.

3. The gait automatic assessment method for cerebral palsy patients according to claim 1, wherein, The gait test: Before the test, the participant first walks habitually on the walking path to adapt to the environment; during the formal test, the participant takes off shoes and walks at a normal speed along a straight track, collecting videos of walking from left to right or from right to left; ensure that the complete body of the participant is captured during the video shooting, and the video needs to collect data of at least 3 gait cycles.

4. The gait automatic evaluation method for cerebral palsy patients according to claim 1, wherein The pose estimation and key point detection: 1) Pose estimation algorithm: The MediaPipe model automatically detects key points in the whole body movement of the patient based on the RGB images captured by the camera; These key points are represented by three-dimensional coordinates of x, y, and z. The x and y coordinates represent the position of the key point in the plane, and the z coordinate is used to estimate the depth of the key point; through the pose estimation algorithm, the system can real-time track the movement trajectory of the patient's lower limbs; 2) Data preprocessing: After obtaining the key point coordinate data, the system preprocesses the original data; this step includes removing noise and filtering abnormal data points; to avoid interference from camera jitter or light changes on the detection results, the system smooths the coordinate changes between frames.

5. The gait automatic evaluation method for cerebral palsy patients according to claim 1, wherein, The gait parameter analysis: 1) Spatiotemporal parameter analysis: The system calculates gait spatiotemporal parameters based on the position information obtained from the posture assessment, such as: walking speed, cadence, step length; 2) Joint angle analysis: The system tracks and analyzes the joint angles of the hip, knee, and ankle; by calculating the angle changes of the joints within the sub - phases of the gait cycle; Gait abnormality diagnosis: Abnormality recognition: The system compares the change curves of joint angles in the sub - phases of the gait cycle with the normal population dataset to identify typical abnormalities of cerebral palsy under single joints; Result output and report generation: Report content: The report details the patient's spatiotemporal parameter information in tabular form, facilitating the therapist's comprehensive assessment of the walking function; the evaluation results of joint angles are presented as curve graphs, intuitively showing the patient's movement characteristics, and are accompanied by a comparison with the normal gait curve to help quickly identify potential abnormalities; the gait abnormalities detected by the automatic algorithm and their possible causes are supplemented with clear explanations, providing an objective reference for the therapist and assisting in formulating more precise rehabilitation strategies; Data comparison and tracking: Supports the comparison of evaluation results at different time periods to help doctors track the progress of patients during the rehabilitation process; by comparing evaluation reports on different dates; Cloud deployment and system maintenance: Cloud deployment: The computing and storage resources of the system are hosted in the cloud, capable of processing the data of multiple patients simultaneously; users only need to upload videos, and the cloud system will automatically analyze and return the evaluation report, with simple operation, suitable for rapid application in clinical environments; System maintenance and update: The system supports later maintenance and update and can be continuously updated with the development of the posture recognition algorithm.

6. A gait automatic evaluation system for cerebral palsy patients, which implements the gait automatic evaluation method for cerebral palsy patients as described in any one of claims 1-5, characterized in that, The gait automatic evaluation system for cerebral palsy patients includes: Data acquisition module, which is responsible for recording the patient's walking process through the smartphone camera in the first step of system operation; the patient needs to walk on a prescribed path; the walking task can be carried out in a hospital, rehabilitation center, or the patient's daily living environment; Detection module, used for pose estimation and key point detection; after data acquisition is completed, the system enters the pose estimation and key point detection module; using the open - source MediaPipe computer vision algorithm, it automatically detects and tracks the movement of the patient's lower limbs from the video; it identifies 33 key points of the human body through a convolutional neural network (CNN), including parts related to the movement of the hip, knee, ankle, foot, and lower limbs; Division module, used for gait cycle division; after pose estimation is completed and key point data is obtained, the system enters the gait cycle division module; it extracts the complete gait cycle based on the identified gait events for subsequent analysis; Analysis module, used for gait parameter analysis; after identifying gait events and obtaining the gait cycle, the system enters the gait parameter analysis module; it calculates gait spatiotemporal parameters and kinematic parameters; these characteristic parameters provide a reference for the therapist to evaluate the patient's walking performance; Diagnosis module, used for gait abnormality diagnosis; after obtaining gait parameters, the system enters the gait abnormality diagnosis module; automatically identifies gait abnormalities and recommends possible causes; this information provides a reference for therapists to analyze the results of gait parameter diagnosis. Output module, used for result output and report generation; after completing the abnormal gait diagnosis, the system enters the report generation module to provide detailed evaluation results for clinicians; the evaluation report includes the joint angle change curve during walking, spatio-temporal parameter results, and abnormal diagnosis and possible causes of joint angles. Maintenance module, used for cloud deployment and system maintenance.

7. A computer device, characterized in that, The computer device includes a memory and a processor. When the computer program stored in the memory is executed by the processor, the processor executes the steps of the gait automatic evaluation method for cerebral palsy patients according to any one of claims 1-5.

8. A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the processor executes the steps of the gait automatic evaluation method for cerebral palsy patients according to any one of claims 1-5.

9. An information data processing terminal, characterized in that, The information data processing terminal is used to implement the gait automatic evaluation system for cerebral palsy patients according to claim 6.

Images