A method and system for motion gait analysis in a telemedicine scenario

Abstract

The application provides a motion gait analysis method and system for a remote medical scene, comprising: acquiring human video image data through an image acquisition device and storing the human video image data in a data storage end; acquiring the human video image data from the data storage end, performing image feature extraction through a preset image feature extraction model, and generating joint coordinate information; matching the joint coordinate information with body size, performing calculation through a preset scale feature extraction model, and generating joint size coordinate information similar to the real human body size; performing deep fusion of the joint coordinate information and the joint size coordinate information through a preset feature fusion model to obtain gait fusion data; and displaying gait fusion results on a display device at a user end through decoding of the gait fusion data. The application solves the problems of low automation, high cost and great limitations of existing gait analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of auxiliary medical technology, in particular to a motion gait analysis method and system for a remote medical scene. BACKGROUND

[0002] The correlation between gait characteristics and cognition has been repeatedly demonstrated in clinical experiments and theoretical analysis. Damage to the nervous system can cause patients to lose concentration, which further leads to obvious abnormalities in their gait, such as slow walking speed and unstable gait. Studies have shown that in the elderly population, if cognitive impairment occurs, the probability of falling is twice that of normal elderly people. In addition, nerve damage can also cause patients to suffer from chronic pain and muscle atrophy. These symptoms limit the patient's normal activities and reduce the patient's quality of life. Gait characteristics are a widely used means of assessing the condition of the nervous system in patients with cognitive impairment. In conditions such as stroke, Parkinson's syndrome, cerebral palsy, arthritis, and muscle atrophy, gait analysis results provide a clinical reference for medical personnel and test subjects. However, existing gait analysis methods mostly rely on expensive experimental equipment and professional medical personnel, which limits the widespread application of existing gait analysis methods, especially in remote areas where resources are scarce and medical resources are insufficient. Therefore, if a gait analysis system for a remote medical scene can be developed to meet the needs of the general public in remote areas for the diagnosis and treatment of neurological diseases and cognitive impairment, it will greatly facilitate the diagnosis and treatment of patients. A gait analysis system in a remote scene needs to be automated, inexpensive, and easy to deploy.

[0003] In existing technical solutions, a reflective marker ball is often used to mark the subject's joint, and the position of the subject's joint in the video is located by the marker to analyze the gait, achieving semi-automated gait analysis. Or a method of intelligent auxiliary analysis of the ankle joint, which can achieve auxiliary diagnosis and treatment of patients, cannot meet the multiple requirements of automation, convenience, and inexpensiveness of a remote gait analysis system. SUMMARY

[0004] The present application provides a motion gait analysis method and system for a remote medical scene to solve the problem of low automation, high cost, and great limitations of existing gait analysis.

[0005] The present application provides a motion gait analysis method for a remote medical scene, comprising:

[0006] Acquiring human video image data through an image acquisition device and storing it in a data storage end;

[0007] Human video image data is obtained from the data storage terminal, and image features are extracted using a preset image feature extraction model to generate joint point coordinate information.

[0008] The joint coordinate information is matched with the body size, and the joint size coordinate information that is similar to the actual human body size is generated by calculating through a preset scale feature extraction model.

[0009] The joint point coordinate information and the joint size coordinate information are deeply fused using a preset feature fusion model to obtain gait fusion data;

[0010] The gait fusion data is decoded and the gait fusion results are displayed on the user's display device.

[0011] According to the present invention, a method for gait analysis in a telemedicine scenario, wherein acquiring human video image data through an image acquisition device and storing it in a data storage terminal specifically includes:

[0012] The image acquisition device captures video of the human body from multiple angles, generating human body video image data;

[0013] The human body video image data is transmitted to a data storage terminal via a wireless network for storage.

[0014] According to the present invention, a gait analysis method for remote medical scenarios is provided, which acquires human video image data from the data storage terminal, extracts image features through a preset image feature extraction model, and generates joint coordinate information, specifically including:

[0015] The image feature extraction model extracts the coordinates of key points related to gait analysis from human video image data;

[0016] The extracted areas include: left and right buttocks, left and right knees, left and right ankles, and left and right toes. Based on the coordinate information of the extracted areas, the angles of the left and right knees, the angles of the left and right ankles, the distance between the left and right ankles, the distance between the left and right toes, and the distance between the left and right knees are calculated.

[0017] According to the present invention, a gait analysis method for remote medical scenarios matches the joint coordinate information with body dimensions, calculates the joint size coordinate information similar to the actual human body size using a preset scale feature extraction model, and specifically includes:

[0018] The scale feature extraction model uses the measured leg length and the coordinate information of the joint points in the image to recover joint information similar to the real scale.

[0019] Define a new joint point as the coordinate center of the scale feature, and obtain the joint size coordinate information with scale constraints based on the scaling ratio and image coordinates.

[0020] According to the present invention, a gait analysis method for remote medical scenarios involves deep fusing the joint point coordinate information and the joint size coordinate information using a preset feature fusion model to obtain gait fusion data, specifically including:

[0021] The joint coordinate information extracted by the image feature extraction model and the joint size coordinate information extracted by the scale feature extraction model are fused together using a feature fusion model.

[0022] The feature fusion model uses a long short-term memory network to process the sequence feature points, and then performs further processing through one-dimensional convolution. The processed deep features are then fused using an attention mechanism to generate gait fusion data.

[0023] According to the present invention, a gait analysis method for remote medical scenarios is provided, wherein the gait fusion data is decoded and the gait fusion result is displayed on a display device at the user end, specifically including:

[0024] The gait fusion data is decoded and output through two fully connected layers to obtain speed, step frequency, and symmetrical gait information;

[0025] The corresponding gait information is displayed on the user's device, allowing doctors to make diagnoses or patients to self-assess their condition.

[0026] The present invention also provides a gait analysis system for remote medical scenarios, the system comprising:

[0027] The image acquisition module acquires human video image data through an image acquisition device and stores it in the data storage terminal;

[0028] The image feature extraction module is used to acquire human video image data from the data storage terminal, extract image features through a preset image feature extraction model, and generate joint point coordinate information.

[0029] The scale feature extraction module matches the joint coordinate information with the body size and calculates it using a preset scale feature extraction model to generate joint size coordinate information similar to the actual human body size.

[0030] The fusion module is used to perform deep fusion of the joint point coordinate information and the joint size coordinate information through a preset feature fusion model to obtain gait fusion data.

[0031] The display module is used to decode the gait fusion data and display the gait fusion results on the user's display device.

[0032] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the motion gait analysis method for telemedicine scenarios as described above.

[0033] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the motion gait analysis method for remote medical scenarios as described above.

[0034] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the motion gait analysis method for remote medical scenarios as described above.

[0035] This invention provides a method and system for gait analysis in telemedicine scenarios. By using only a regular monocular camera at the user end to collect the patient's body posture information without requiring special environmental deployment, and by deploying an automated analysis module at the diagnosis end, automated gait analysis and diagnosis of patients can be achieved. This enables remote doctor diagnosis or patient self-diagnosis, reduces costs, expands application scenarios, and provides high-quality medical resources for patients in remote areas with insufficient medical resources. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0037] Figure 1 This is one of the flowcharts of a motion gait analysis method for remote medical scenarios provided by the present invention;

[0038] Figure 2 This is the second flowchart of a method for motion gait analysis in a telemedicine scenario provided by the present invention;

[0039] Figure 3 This is the third flowchart of a method for motion gait analysis in a telemedicine scenario provided by the present invention;

[0040] Figure 4This is the fourth flowchart of a method for motion gait analysis in a telemedicine scenario provided by the present invention;

[0041] Figure 5 This is the fifth flowchart of a method for motion gait analysis in a telemedicine scenario provided by the present invention;

[0042] Figure 6 This is the sixth flowchart of a method for motion gait analysis in a telemedicine scenario provided by the present invention;

[0043] Figure 7 This is a schematic diagram of the module connection of a motion gait analysis system for remote medical scenarios provided by the present invention;

[0044] Figure 8 This is a schematic diagram of the structure of the electronic device provided by the present invention.

[0045] Figure label:

[0046] 110: Image acquisition module; 120: Image feature extraction module; 130: Scale feature extraction module; 140: Fusion module; 150: Display module;

[0047] 810: Processor; 820: Communication interface; 830: Memory; 840: Communication bus. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0049] The following is combined Figures 1-6 This invention describes a gait analysis method for use in a telemedicine scenario, comprising:

[0050] S100: Acquire human video image data through an image acquisition device and store it in a data storage terminal;

[0051] S200: Obtain human video image data from the data storage terminal, extract image features through a preset image feature extraction model, and generate joint point coordinate information;

[0052] S300. Match the joint coordinate information with the body size, and calculate using a preset scale feature extraction model to generate joint size coordinate information similar to the actual human body size.

[0053] S400. The joint point coordinate information and the joint size coordinate information are deeply fused using a preset feature fusion model to obtain gait fusion data;

[0054] S500: The gait fusion data is decoded and the gait fusion result is displayed on the user's display device.

[0055] This invention can acquire human video images using ordinary video capture devices or mobile terminal cameras, making it applicable to a wide range of scenarios and with a low barrier to entry. After processing the human video image data, gait fusion data can be established. After parsing the gait fusion data, the gait information can be displayed through a display device, which can quickly analyze the human body's physical condition and facilitate more patients to access relevant medical resources at low cost.

[0056] Human body video image data is acquired through image acquisition devices and stored in a data storage terminal, specifically including:

[0057] S101. The image acquisition device captures video of the human body from multiple angles and generates human body video image data.

[0058] S102, The human body video image data is transmitted to the data storage terminal via a wireless network for storage.

[0059] In this invention, the image acquisition device can use a common monocular camera or a mobile terminal with a camera function, such as a mobile phone or tablet, to capture video. After acquisition, the video is transmitted to a data storage terminal via a wireless network. Upon receiving the data, the data storage terminal backs up the human body video image data and then transmits it to the diagnostic terminal via a wireless network. Using common image acquisition devices reduces costs and allows more people to use gait analysis for diagnosis and treatment.

[0060] Human video image data is acquired from the data storage terminal, and image features are extracted using a preset image feature extraction model to generate joint point coordinate information, specifically including:

[0061] S201, The image feature extraction model extracts the coordinates of key points related to gait analysis in the image from human video image data;

[0062] S202. The extracted areas include: left and right buttocks, left and right knees, left and right ankles, and left and right toes. The angles of the left and right knees, the angles of the left and right ankles, the distance between the left and right ankles, the distance between the left and right toes, and the distance between the left and right knees are calculated based on the coordinate information of the extracted areas.

[0063] In this invention, an automatic gait analysis algorithm is applied at the diagnostic end for gait analysis. The image feature extraction model uses the well-known pose analysis algorithm OpenPose to extract the joint coordinates of the subject in the image. Here, only the joint coordinates related to gait analysis are extracted, including: left and right hips, left and right knees, left and right ankles, and left and right toes. Using these coordinates, seven gait-related features are further calculated: the angles of the left and right knees, the angles of the left and right ankles, the distance between the left and right ankles, the distance between the left and right toes, and the distance between the left and right knees, generating joint coordinate information.

[0064] The joint coordinate information is matched with body dimensions, and calculations are performed using a preset scale feature extraction model to generate joint dimension coordinate information similar to real human body dimensions, specifically including:

[0065] S301, The scale feature extraction model uses the measured leg length and the coordinate information of the joint points in the image to recover joint information similar to the real scale;

[0066] S302. Define a new joint point as the coordinate center of the scale feature. Based on the scaling ratio and image coordinates, the joint size coordinate information with scale constraints can be obtained.

[0067] In this invention, the scale feature extraction model uses measured leg length and the coordinate information of joints in the image to recover joint information similar to the real scale. First, a new joint point, called the hip center, is defined as the coordinate center for the scale feature:

[0068]

[0069] Where (x) mh ,y mh ), (x rh ,y rh ) and (x lh ,y lh These represent the coordinates of the center, left, and right sides of the buttocks in the image, respectively. Then, the ratios of the thigh to the calf of each leg are calculated:

[0070]

[0071]

[0072] Where (x) rk ,y rk ),(x lk ,y lk ), (x ra ,y ra ), (x la ,y la) represent the coordinates of the left and right knees and the left and right ankles in the image, respectively; α l α r These represent the ratios between the left and right thighs and calves, respectively; n is the total number of frames in the video stream; and i is the corresponding video frame. Considering that the left and right legs are symmetrical:

[0073]

[0074] The average ratio between the left and right thighs and calves is taken as the thigh-to-calf ratio for the subject in this sequence. Based on this ratio and the measured leg length *l*, the thigh length and calf length can be calculated.

[0075]

[0076]

[0077] Among them l t It is the length of the thigh, l c This refers to the length of the lower leg. According to the pinhole camera model, there is a linear mapping between image coordinates and real-world coordinates. This scaling factor is determined by the distance between the subject and the camera, as well as the camera parameters. Considering that gait analysis videos are generally filmed from the side of the subject, the distances from the left and right legs to the camera are different. Therefore, two different sets of scaling parameters are used to scale the subject's left and right legs respectively:

[0078]

[0079]

[0080]

[0081]

[0082] in It is the scaling ratio of the right leg along the x and y axes; where This refers to the scaling ratio of the left leg along the x and y axes. Since there are n frames in the same video sequence, there are a total of 4n constraints. The least squares method is used to solve this overdetermined problem. Based on the solved scaling ratio and image coordinates, coordinate information with scale constraints can be obtained:

[0083]

[0084]

[0085]

[0086]

[0087] in These represent the coordinates of the left and right knees and ankles in reality after scaling.

[0088] The joint point coordinate information and the joint size coordinate information are deeply fused using a preset feature fusion model to obtain gait fusion data, specifically including:

[0089] S401. The joint coordinate information extracted by the image feature extraction model and the joint size coordinate information extracted by the scale feature extraction model are fused together using a feature fusion model.

[0090] S402. The feature fusion model uses a long short-term memory network to process the sequence feature points, and then performs further processing through one-dimensional convolution. The processed deep features are fused using an attention mechanism to generate gait fusion data.

[0091] Gait fusion data is decoded and the gait fusion results are displayed on the user's display device, specifically including:

[0092] S501, the gait fusion data is decoded and output through two fully connected layers to obtain speed, step frequency, and symmetrical gait information;

[0093] S502: Display the corresponding gait information on the user's display device so that doctors can make diagnoses or patients can conduct self-assessments of their condition.

[0094] On the diagnostic and treatment side, an automated analysis system is used for diagnostic and treatment analysis. When abnormalities occur, the data with abnormal results from the automated analysis are further analyzed manually to improve the accuracy of the analysis.

[0095] This invention provides a gait analysis method for telemedicine scenarios. By using only a regular monocular camera at the user end to collect the patient's body posture information without requiring special environmental deployment, and by deploying an automated analysis module at the diagnosis end, automated gait analysis and diagnosis of patients can be achieved. This enables remote doctor diagnosis or patient self-diagnosis, reduces costs, expands application scenarios, and provides high-quality medical resources for patients in remote areas with insufficient medical resources.

[0096] refer to Figure 7 The present invention also discloses a gait analysis system for remote medical scenarios, the system comprising:

[0097] The image acquisition module 110 acquires human video image data through the image acquisition device and stores it in the data storage terminal;

[0098] Image feature extraction module 120 is used to acquire human video image data from the data storage terminal, extract image features through a preset image feature extraction model, and generate joint point coordinate information;

[0099] The scale feature extraction module 130 matches the joint coordinate information with the body size and calculates it using a preset scale feature extraction model to generate joint size coordinate information similar to the real human body size.

[0100] The fusion module 140 is used to perform deep fusion of the joint point coordinate information and the joint size coordinate information through a preset feature fusion model to obtain gait fusion data;

[0101] The display module 150 is used to display the gait fusion results on the user's display device after the gait fusion data is decoded.

[0102] The image acquisition module 110 captures video of the human body from multiple angles using the image acquisition device, generating human body video image data.

[0103] The human body video image data is transmitted to a data storage terminal via a wireless network for storage.

[0104] Image feature extraction module 120 extracts the coordinates of key points related to gait analysis from human video image data using the image feature extraction model;

[0105] The extracted areas include: left and right buttocks, left and right knees, left and right ankles, and left and right toes. Based on the coordinate information of the extracted areas, the angles of the left and right knees, the angles of the left and right ankles, the distance between the left and right ankles, the distance between the left and right toes, and the distance between the left and right knees are calculated.

[0106] The scale feature extraction module 130 uses the measured leg length and the coordinate information of the joint points in the image to recover joint information similar to the real scale through the scale feature extraction model.

[0107] Define a new joint point as the coordinate center of the scale feature, and obtain the joint size coordinate information with scale constraints based on the scaling ratio and image coordinates.

[0108] The fusion module 140 fuses the joint coordinate information extracted by the image feature extraction model with the joint size coordinate information extracted by the scale feature extraction model through a feature fusion model.

[0109] The feature fusion model uses a long short-term memory network to process the sequence feature points, and then performs further processing through one-dimensional convolution. The processed deep features are then fused using an attention mechanism to generate gait fusion data.

[0110] The display module 150 decodes the gait fusion data through two fully connected layers to output speed, cadence, and symmetrical gait information;

[0111] The corresponding gait information is displayed on the user's device, allowing doctors to make diagnoses or patients to self-assess their condition.

[0112] This invention provides a gait analysis system for telemedicine scenarios. By using only a regular monocular camera at the user end, it can collect the patient's body posture information without requiring special environmental deployment. At the diagnosis and treatment end, an automated analysis module can be deployed to achieve automated gait analysis and diagnosis of patients, enabling remote doctor diagnosis or patient self-diagnosis, reducing costs, expanding application scenarios, and providing high-quality medical resources for patients in remote areas with insufficient medical resources.

[0113] Figure 8 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 8 As shown, the electronic device may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840. The processor 810, communications interface 820, and memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute a gait analysis method for remote medical scenarios. This method includes: acquiring human video image data through an image acquisition device and storing it in a data storage terminal.

[0114] Human video image data is obtained from the data storage terminal, and image features are extracted using a preset image feature extraction model to generate joint point coordinate information.

[0115] The joint coordinate information is matched with the body size, and the joint size coordinate information that is similar to the actual human body size is generated by calculating through a preset scale feature extraction model.

[0116] The joint point coordinate information and the joint size coordinate information are deeply fused using a preset feature fusion model to obtain gait fusion data;

[0117] The gait fusion data is decoded and the gait fusion results are displayed on the user's display device.

[0118] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0119] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program, the computer program being stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer is able to execute a motion gait analysis method for remote medical scenarios provided by the above methods, the method including: acquiring human video image data through an image acquisition device and storing it in a data storage terminal;

[0120] Human video image data is obtained from the data storage terminal, and image features are extracted using a preset image feature extraction model to generate joint point coordinate information.

[0121] The joint coordinate information is matched with the body size, and the joint size coordinate information that is similar to the actual human body size is generated by calculating through a preset scale feature extraction model.

[0122] The joint point coordinate information and the joint size coordinate information are deeply fused using a preset feature fusion model to obtain gait fusion data;

[0123] The gait fusion data is decoded and the gait fusion results are displayed on the user's display device.

[0124] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform a motion gait analysis method for remote medical scenarios provided by the methods described above, the method comprising: acquiring human video image data through an image acquisition device and storing it in a data storage terminal;

[0125] Human video image data is obtained from the data storage terminal, and image features are extracted using a preset image feature extraction model to generate joint point coordinate information.

[0126] The joint coordinate information is matched with the body size, and the joint size coordinate information that is similar to the actual human body size is generated by calculating through a preset scale feature extraction model.

[0127] The joint point coordinate information and the joint size coordinate information are deeply fused using a preset feature fusion model to obtain gait fusion data;

[0128] The gait fusion data is decoded and the gait fusion results are displayed on the user's display device.

[0129] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0130] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for gait analysis in telemedicine scenarios, characterized in that, include: Human body video image data is acquired through image acquisition equipment and stored in the data storage terminal; Human video image data is obtained from the data storage terminal, and image features are extracted using a preset image feature extraction model to generate joint point coordinate information. The joint coordinate information is matched with the body size, and the joint size coordinate information that is similar to the actual human body size is generated by calculating through a preset scale feature extraction model. The joint coordinate information is matched with body dimensions, and calculations are performed using a preset scale feature extraction model to generate joint dimension coordinate information similar to real human body dimensions, specifically including: The scale feature extraction model uses the measured leg length and the coordinate information of the joint points in the image to recover joint information similar to the real scale. Define a new joint point as the coordinate center of the scale feature, and obtain the joint size coordinate information with scale constraints based on the scaling ratio and image coordinates. The scale feature extraction model uses the measured leg length and the coordinate information of the joints in the image to recover joint information similar to the real scale. First, a new joint point is defined as the hip center, which is used as the coordinate center of the scale feature: Where (x) mh ,y mh ), (x rh ,y rh ) and (x lh ,y lh ) represent the coordinates of the center, left side, and right side of the buttocks in the image, respectively; Then, the ratio of the thigh to the calf of each leg was calculated separately: Where (x) rk ,y rk ), (x lk ,y lk ), (x ra ,y ra ), (x la ,y la ) represent the coordinates of the left and right knees and the left and right ankles in the image, respectively; α l α r These represent the ratios between the left and right thighs and calves, respectively; n is the total number of frames in the video stream; and i is the video of the corresponding frame. Considering that the human left and right legs are symmetrical: The average ratio between the left and right thighs and calves is taken as the subject's thigh-to-calf ratio. Based on this ratio and the measured leg length l, the thigh length and calf length are calculated. Among them l t It is the length of the thigh, l c It is the length of the lower leg; According to the pinhole imaging model, there is a linear mapping relationship between image coordinates and real-world coordinates. This scaling factor is determined by the distance between the subject and the camera, as well as the camera parameters. Considering that the gait analysis video is taken from the side of the subject, the distances from the subject's left and right legs to the camera are different. Therefore, two different sets of scaling parameters are used to scale the subject's left and right legs respectively. in It is the scaling ratio of the right leg along the x and y axes; This refers to the scaling ratio of the left leg along the x and y axes; there are n frames in the same video set, therefore there are 4n sets of constraints. The least squares method is used to solve the overdetermined problem; based on the solved scaling ratio and image coordinates, coordinate information with scale constraints is obtained: in These represent the coordinates of the left and right knees and the left and right ankles in reality after scaling. The joint point coordinate information and the joint size coordinate information are deeply fused using a preset feature fusion model to obtain gait fusion data; The gait fusion data is decoded and the gait fusion results are displayed on the user's display device.

2. The gait analysis method for telemedicine scenarios according to claim 1, characterized in that, The acquisition of human video image data through an image acquisition device and its storage in a data storage terminal specifically includes: The image acquisition device captures video of the human body from multiple angles, generating human body video image data; The human body video image data is transmitted to a data storage terminal via a wireless network for storage.

3. The gait analysis method for telemedicine scenarios according to claim 1, characterized in that, Human video image data is acquired from the data storage terminal, and image features are extracted using a preset image feature extraction model to generate joint point coordinate information, specifically including: The image feature extraction model extracts the coordinates of key points related to gait analysis from human video image data; The extracted areas include: left and right buttocks, left and right knees, left and right ankles, and left and right toes. Based on the coordinate information of the extracted areas, the angles of the left and right knees, the angles of the left and right ankles, the distance between the left and right ankles, the distance between the left and right toes, and the distance between the left and right knees are calculated.

4. The gait analysis method for telemedicine scenarios according to claim 1, characterized in that, The joint point coordinate information and the joint size coordinate information are deeply fused using a preset feature fusion model to obtain gait fusion data, specifically including: The joint coordinate information extracted by the image feature extraction model and the joint size coordinate information extracted by the scale feature extraction model are fused together using a feature fusion model. The feature fusion model uses a long short-term memory network to process the sequence feature points, and then performs further processing through one-dimensional convolution. The processed deep features are then fused using an attention mechanism to generate gait fusion data.

5. The gait analysis method for telemedicine scenarios according to claim 1, characterized in that, The gait fusion data is decoded, and the gait fusion results are displayed on the user's display device, specifically including: The gait fusion data is decoded and output through two fully connected layers to obtain speed, step frequency, and symmetrical gait information; The corresponding gait information is displayed on the user's device, allowing doctors to make diagnoses or patients to self-assess their condition.

6. A gait analysis system for telemedicine scenarios, employing the gait analysis method as described in any one of claims 1 to 5, characterized in that, The system includes: The image acquisition module acquires human video image data through an image acquisition device and stores it in the data storage terminal; The image feature extraction module is used to acquire human video image data from the data storage terminal, extract image features through a preset image feature extraction model, and generate joint point coordinate information. The scale feature extraction module matches the joint coordinate information with the body size and calculates it using a preset scale feature extraction model to generate joint size coordinate information similar to the actual human body size. The fusion module is used to perform deep fusion of the joint point coordinate information and the joint size coordinate information through a preset feature fusion model to obtain gait fusion data. The display module is used to decode the gait fusion data and display the gait fusion results on the user's display device.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the gait analysis method for remote medical scenarios as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the gait analysis method for remote medical scenarios as described in any one of claims 1 to 5.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the gait analysis method for remote medical scenarios as described in any one of claims 1 to 5.

Images