Ankle abnormity intelligent identification method based on three-point detection method

Abstract

The invention discloses an intelligent ankle anomaly identification method based on a three-point detection method. The method comprises the steps of S1, data acquisition; s2, key point definition; s3, key point detection; and S4, feature calculation and evaluation. And calculating varus, valgus and other abnormal states of the ankle by identifying three key points, namely a shank midpoint, a medial and lateral ankle midpoint and a heel midpoint. The method combines an image technology and an artificial intelligence algorithm, realizes quantitative evaluation of ankle health, has the characteristics of non-contact, low cost and high efficiency, and is widely applied to the fields of medical health, exercise rehabilitation and shoe design.

Description

Technical Field

[0001] The present invention relates to the technical fields of medical image processing and artificial intelligence, and particularly relates to an intelligent foot and ankle abnormality recognition method based on a three-point detection method. Background Art

[0002] The foot and ankle are important parts of the human body, bearing the functions of supporting body weight, balancing the body, and coordinating movements. However, foot and ankle problems are very common, affecting people's quality of life. Foot and ankle abnormalities (such as varus and valgus) can affect the gait balance and motor function of the human body, and may lead to diseases of the knee, hip joint, and even lumbar spine. Traditional detection methods rely on manual observation or complex equipment, with problems of strong subjectivity and complex operation. By identifying three key points of the foot and ankle and calculating the relative positional relationship, automatic detection and quantitative analysis of the foot and ankle state can be achieved. Traditional foot and ankle detection methods mainly rely on manual observation and palpation, with problems of strong subjectivity and low efficiency. With the development of technology, some new detection methods have gradually emerged, such as X-ray, CT, MRI, etc. These methods can provide more objective and detailed foot and ankle information, but have disadvantages such as high cost and large radiation.

[0003] Therefore, it is necessary to provide an intelligent foot and ankle abnormality recognition method based on a three-point detection method to solve the above problems. The present invention realizes automatic and quantitative evaluation of the foot and ankle state by accurately identifying three key points of the foot and ankle and calculating the foot and ankle angles and related parameters. Summary of the Invention

[0004] The purpose of this part is to outline some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Simplifications or omissions may be made in this part, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this part, the abstract, and the title. However, such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] To solve the above technical problems, the present invention provides the following technical solution: An intelligent foot and ankle abnormality recognition method based on a three-point detection method, characterized by including:

[0006] S1. Data acquisition, using an RGB camera or a depth camera to acquire images of the front and rear of the foot and ankle, heel, and calf;

[0007] S2. Definition of key points, including the midpoint of the calf, the midpoint of the medial and lateral malleoli, and the midpoint of the heel;

[0008] S3. Detection of key points, using the YOLOv11-pose model;

[0009] S4. Feature calculation and evaluation, calculating the foot and ankle axis angle and evaluating the foot and ankle stability;

[0010] Non-contact detection: Images including the front views of the ankle, heel, and calf are collected through ordinary cameras or depth cameras, without the need for expensive equipment and complex operations.

[0011] Key point recognition: Through imaging technology and artificial intelligence algorithms, three key points, namely the midpoint of the calf, the midpoint of the medial and lateral malleoli, and the midpoint of the heel, are accurately detected.

[0012] Feature calculation: Based on the coordinates of the three points, feature parameters such as ankle angles are calculated to evaluate varus, valgus, and other abnormal states.

[0013] As a preferred embodiment of the intelligent ankle abnormality recognition method based on the three-point detection method of the present invention, in S1, the collected data is processed by robustness processing, and methods such as image enhancement and histogram equalization are used to process images with insufficient light or overexposure to improve image quality.

[0014] As a preferred embodiment of the intelligent ankle abnormality recognition method based on the three-point detection method of the present invention, the midpoint of the calf in S2 is the midpoint of the axes of the tibia and fibula; the midpoint of the medial and lateral malleoli is the midpoint between the medial and lateral malleoli; the midpoint of the heel is the center point of the posterior edge of the calcaneus.

[0015] As a preferred embodiment of the intelligent ankle abnormality recognition method based on the three-point detection method of the present invention, in S3, an ELAN module is further included, and the ELAN module extracts feature information in the image through layer aggregation and feature fusion.

[0016] As a preferred embodiment of the intelligent ankle abnormality recognition method based on the three-point detection method of the present invention, in S3, the YOLOv11-pose model is used, which supports a variety of data augmentation techniques and various training techniques.

[0017] As a preferred embodiment of the intelligent ankle abnormality recognition method based on the three-point detection method of the present invention, in S4, it includes defining coordinate points, calculating vectors, calculating the modulus of vectors, calculating the dot product of vectors, and calculating angles.

[0018] Advantages of the present invention: Precise detection: Through the three-point detection method, the ankle state is quantified to avoid subjective errors; Automatic operation: The entire process from image acquisition to abnormality judgment is automatically completed; Low-cost and high-efficiency: It can be achieved only with ordinary imaging equipment; Applicable to multiple fields: It can be applied in medical, sports, and industrial design. In view of the deficiencies of the prior art, the present invention proposes to accurately identify three key points of the ankle, calculate ankle angles and related parameters, and realize automatic and quantitative evaluation of the ankle state. This method has the characteristics of non-contact, low cost, and high efficiency, effectively solving the deficiencies of traditional detection methods, and has important clinical application value and social significance. Description of the Drawings

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0020] Among them:

[0021] Figure 1 It is a step flowchart of an intelligent foot and ankle abnormality recognition method based on the three-point detection method according to an embodiment provided by the present invention;

[0022] Figure 2 It is a schematic diagram of three points of the foot and ankle of an intelligent foot and ankle abnormality recognition method based on the three-point detection method according to an embodiment provided by the present invention. Specific Embodiments

[0023] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will make a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments.

[0024] Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0026] Embodiment 1

[0027] Referring to Figure 1-2 , the first embodiment of the present invention, an intelligent foot and ankle abnormality recognition method based on the three-point detection method, characterized by including:

[0028] S1. Data collection, using an RGB camera or a depth camera to collect images of the foot and ankle, heel, and the back of the calf;

[0029] S2. Key point definition, including the midpoint of the calf, the midpoint of the inner and outer ankles, and the midpoint of the heel;

[0030] S3. Key point detection, using the YOLOv11-pose model;

[0031] S4. Feature calculation and evaluation, calculating the angle of the foot and ankle axis and evaluating the stability of the foot and ankle;

[0032] Specifically, the device used for S1 data acquisition is an RGB camera or a depth camera to collect images of the ankle, heel, and the back of the calf. The depth camera can provide depth information, which helps to more accurately identify the ankle structure. At the same time, the acquisition can be carried out according to the different postures of the subjects. The subjects stand barefoot with their legs parallel and the center of gravity evenly distributed, and the camera is perpendicular to the ankle area for shooting. The shooting distance should be kept consistent to reduce the impact of perspective distortion on the results. Robust processing is performed on the collected data. In terms of the background, when processing, try to select an environment with a simple background to reduce the interference environment; in terms of the shooting angle, calibrate the image distortion caused by the shooting angle through a calibration object or a reference plane; in terms of the lighting conditions, use methods such as image enhancement and histogram equalization to process images with insufficient lighting or overexposure to improve the image quality. S2, Key point definition, Midpoint of the Lower Leg: It is defined as the midpoint of the axes of the tibia and fibula, usually located in the lower 1 / 3 area of the calf; the role of this definition is to serve as a longitudinal reference point for the tibia and is used to judge the deviation of the overall axis of the ankle. The defined method is: manually or automatically calibrate the edge lines of the tibia and fibula in the image; calculate the midline of the two edge lines, which is the center line of the calf; take a point 1 / 3 down along the center line, which is the midpoint of the lower leg. Midpoint between Malleoli in S2: It is defined as the midpoint between the medial malleolus and the lateral malleolus. The role of the midpoint between the malleoli is to provide a lateral reference point for the ankle joint and evaluate the varus and valgus conditions of the ankle. The defined method used is to manually or automatically calibrate the highest points of the medial malleolus and the lateral malleolus in the image, and connect the two points and take the midpoint, which is the midpoint between the malleoli. Midpoint of the Heel in S2: It is defined as the center point of the posterior edge of the calcaneus. The role of the midpoint of the heel, as a reference point at the back of the foot, is used to calculate the angle between the ankle axis and the mid-axis of the calf. The defined method used is to manually or automatically calibrate the contour line of the posterior edge of the calcaneus in the image and calculate the center point of the contour line, which is the midpoint of the heel. In S3, key point detection, the YOLOv11-pose model algorithm is used. YOLOv11-pose is the latest object detection model in the YOLO (You Only Look Once) series, and it has achieved significant improvements in terms of speed and accuracy. YOLOv11-pose adopts a new backbone network and decoupled head design, which can better handle objects of different scales and has a faster inference speed. In addition, YOLOv11-pose also supports a variety of data augmentation and training techniques, which can further improve the performance of the model. Among them, YOLOv11-pose adopts a new backbone network called ELAN (Efficient Layer Aggregation Network).The ELAN module can extract feature information in images more effectively through efficient layer aggregation and feature fusion. At the same time, YOLOv11-pose uses a decoupled head design to separate the classification and regression tasks. This design can improve the accuracy of the model and can better handle targets of different scales. In addition, YOLOv11-pose adds a pose estimation head for predicting the coordinates of human key points. The pose estimation head usually consists of multiple convolutional layers and can learn the spatial relationships of human key points. To improve the performance of the model, YOLOv11-pose uses a new loss function called VFL (Varifocal Loss). The VFL loss function can better align the classification and regression tasks, thus improving the performance of the model. YOLOv11-pose supports a variety of data augmentation techniques, including Mosaic, MixUp, CutMix, etc. These techniques can increase the diversity of data, thereby improving the robustness of the model. In the present invention, YOLOv11-pose also supports a variety of training techniques, including EMA (Exponential Moving Average), Warmup, etc. The above techniques can improve the convergence speed and stability of the model.

[0033] The training process is as follows: 1. Data annotation: Use tools such as LabelMe to annotate the smallest rectangular box containing the complete human body in the image; within the bounding box, annotate the coordinates of human key points, such as the midpoint of the calf, the midpoint of the inner and outer ankles, and the midpoint of the heel; for key points that are occluded or outside the image range, they can be not annotated or marked; the annotation format can be COCO, etc. 2. Data preprocessing: Convert the annotated data into the format required by the YOLOv11-pose model; perform operations such as scaling and cropping on the image to adapt to the model input size. 3. Model training: Use the annotated dataset to train the YOLOv11-pose model; use a pre-trained model for fine-tuning to speed up the training speed; adjust model parameters, such as the learning rate, batch size, etc., to achieve the best performance. 4. Model evaluation: Use the test set to evaluate the performance of the model, such as using metrics such as mAP (mean Average Precision); adjust the model parameters according to the evaluation results until the requirements are met.

[0034] S4. Feature calculation and evaluation, ankle axis angle calculation and ankle stability evaluation; the ankle axis angle calculation includes the following steps:

[0035] (1) Define coordinate points

[0036] (x_LegMid, y_LegMid) is the coordinate of the midpoint of the calf;

[0037] (x_InnerOuterAnkle, y_InnerOuterAnkle) are the coordinates of the midpoint of the inner and outer ankles;

[0038] (x_Heel, y_Heel) are the coordinates of the midpoint of the heel;

[0039] (2) Calculate the vector

[0040] The vector v1 from the midpoint of the calf to the midpoint of the inner and outer ankles:

[0041] v1 = (x_InnerOuterAnkle - x_LegMid, y_InnerOuterAnkle - y_LegMid)

[0042] The vector v2 from the midpoint of the inner and outer ankles to the midpoint of the heel:

[0043] v2 = (x_Heel - x_InnerOuterAnkle, y_Heel - y_InnerOuterAnkle)

[0044] (3) Calculate the magnitude of the vector:

[0045] |v1|:

[0046] |v1| = sqrt((x_InnerOuterAnkle - x_LegMid)^2 + (y_InnerOuterAnkle - y_LegMid)^2)

[0047] |v2|:

[0048] |v2| = sqrt((x_Heel - x_InnerOuterAnkle)^2 + (y_Heel - y_InnerOuterAnkle)^2)

[0049] (4) Calculate the dot product of the vectors:

[0050] v1 · v2:

[0051] v1 · v2 = (x_InnerOuterAnkle - x_LegMid) * (x_Heel - x_InnerOuterAnkle) + (y_InnerOuterAnkle - y_LegMid) * (y_Heel - y_InnerOuterAnkle)

[0052] (5) Calculate the included angle θ:

[0053] θ = arccos((v1 · v2) / (|v1| * |v2|))

[0054] (6) Substitute the above calculation results into the formula to obtain:

[0055] θ = arccos(((x_InnerOuterAnkle - x_LegMid)*(x_Heel - x_InnerOuterAnkle)+(y_InnerOuterAnkle - y_LegMid)*(y_Heel

[0056] - y_InnerOuterAnkle)) / (sqrt((x_InnerOuterAnkle - x_LegMid)^2+(y_InnerOuterAnkle - y_LegMid)^2)*sqrt((x_Heel - x_InnerOuterAnkle)^2+(y_Heely_InnerOuterAnkle)^2)))

[0057] (7) Evaluate

[0058] Evaluate the varus and valgus conditions of the ankle according to the angle θ.

[0059] The ankle stability evaluation is as follows:

[0060] By calculating the distance from the midpoint of the inner and outer ankles to the midpoint of the heel, and the distance from the midpoint of the calf to the midpoint of the inner and outer ankles.

[0061] Calculate the ratio of the two distances as an evaluation index for ankle stability; among them, the larger the ratio, the better the ankle stability; the smaller the ratio, the worse the ankle stability.

[0062] For example, common ankle problems include: (1) Ankle deformities: such as varus, valgus, flat feet, high arches, etc. These deformities will affect the normal function of the ankle, resulting in gait abnormalities, easy sprains, etc. (2) Ankle pain: There are many reasons for ankle pain, including sprains, fractures, arthritis, tendinitis, etc. Long-term ankle pain will affect people's daily life and work. (3) Ankle instability: Ankle instability refers to the decline in the stability of the ankle joint, which is prone to sprains. The reasons for ankle instability include ligament injuries, insufficient muscle strength, etc. (4) Ankle diseases: Some diseases, such as diabetes, rheumatoid arthritis, etc., will also cause ankle problems. The application scenarios of the present invention include medical health, sports science, footwear design, etc. Medical health: Assist in diagnosing ankle varus, valgus and other abnormal states, and provide a quantitative evaluation basis for doctors. Sports science: Evaluate the gait and ankle stability of athletes, and optimize the training plan. Footwear design: Provide ankle data support for customized footwear products and improve wearing comfort.

[0063] In summary, through the three-point detection method, the present invention provides an efficient and accurate ankle health assessment method with extensive practical value. By accurately identifying three key points of the ankle, calculating the ankle angles and related parameters, an automated and quantitative assessment of the ankle state is achieved. This method is non-contact, low-cost, and efficient, effectively solving the deficiencies of traditional detection methods, and has important clinical application value and social significance.

[0064] In addition, to provide a concise description of the exemplary embodiments, all features of the actual embodiments may not be described (i.e., those features that are not relevant to the currently contemplated best mode of carrying out the present invention or those features that are not relevant to the implementation of the present invention).

[0065] It should be understood that in the development of any actual implementation, such as in any engineering or design project, numerous specific implementation decisions may be made. Such development efforts may be complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, without undue experimentation, such development efforts will be a routine task of design, manufacturing, and production.

[0066] 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 preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. An intelligent recognition method for foot and ankle abnormalities based on a three-point detection method, characterized in that: include: S1, data collection, using an RGB camera or a depth camera to collect images of the ankle, heel, and calf from the back; S2, definition of key points, including the midpoint of the calf, the midpoint of the inner and outer ankles, and the midpoint of the heel; S3, key point detection, using YOLOv11-pose model; S4, feature calculation and evaluation, ankle axis angle calculation and ankle stability evaluation.

2. The method for intelligently identifying foot and ankle abnormalities based on the three-point detection method according to claim 1 is characterized in that: In S1, the collected data is robustly processed, and image enhancement and histogram equalization methods are used to process images with insufficient light or overexposure to improve image quality.

3. The method for intelligently identifying foot and ankle abnormalities based on the three-point detection method according to claim 1 is characterized in that: The midpoint of the calf in S2 is the midpoint of the axis of the tibia and fibula; the midpoint of the medial and lateral malleolus is the midpoint between the medial and lateral malleolus; and the midpoint of the heel is the center point of the posterior edge of the calcaneus.

4. The method for intelligently identifying foot and ankle abnormalities based on the three-point detection method according to claim 1 is characterized in that: The S3 also includes an ELAN module, which extracts feature information from the image through layer aggregation and feature fusion.

5. The intelligent recognition method for foot and ankle abnormalities based on the three-point detection method according to claim 1 is characterized in that: The YOLOv11-pose model in S3 supports multiple data enhancement technologies and multiple training techniques.

6. The method for intelligently identifying foot and ankle abnormalities based on the three-point detection method according to claim 1 is characterized in that: The S4 includes defining coordinate points, calculating vectors, calculating the modulus of the vectors, calculating the dot product of the vectors, and calculating the angle.

Images