Vision-Based Analysis Of Stroke Rehabilitation Movements

There are a large number of stroke patients in my country,and the number is increasing year by year.Under the current medical conditions in our country,stroke patients need specific limb movement training to restore the corresponding limb function,and the rehabilitation training process for stroke patients is generally carried out at home.The purpose of this study is to use computer vision technology to quantitatively evaluate the rehabilitation effect of stroke patients;at the same time,it can also provide multi-dimensional and data-based evaluation index information for medical staff.Taking the evaluation and analysis of the rehabilitation effect of limb function in stroke patients as the research goal,the main research contents of this paper include:In this paper,the limbs and limb movement characteristics of stroke patients were firstly studied,and the upper limbs of the patients were selected as the research objects.According to the movement characteristics of the patients’ upper limbs,five types of upper limb movements were used as the benchmark to analyze the upper limb rehabilitation movements of stroke patients.Based on these five movements and medical principles,the study defines the corresponding quantifiable evaluation parameters,that is,the maximum angle of raising the hand,the time spent in the process of raising the hand,and the standard degree of the raising of the hand.Corresponding to the acquisition requirements of the evaluation parameters,this paper uses the openpose method to obtain the joint points of the patient’s bones and supplements the missing joint points.Then,according to the obtained upper limb joint p oints and hip joint coordinates,relying on the defined wrist joint lift angle and elbow joint lift angle,design a calculation method,so as to obtain the corresponding angle in each frame in the video stream data of the rehabilitation action,identify and mark the maximum degree among them,calculate the start and end time of the upper limb movement to obtain the duration of the entire training action process,and carry out a complete movement trajectory.track.Based on the tracking and data-based measurement of the movement trajectory,this paper transforms the analysis problem of the quality of rehabilitation exercise into the curve similarity analysis problem of the angle changing with the number of frames during the upper limb raising action.According to the measured value of the corresponding angle in each frame of the video stream,the curve of the angle changing with the number of frames during upper limb training is obtained.For the jitter phenomenon in the curve,the smoothing process is adopted,and then the dynamic warping method is used to analyze the similarity of the ordered curve,so as to calculate the quantitative evaluation information of the action quality.Integrating the above data analysis process,pyQt technology is used to develop a quantitative analysis system for stroke rehabilitation movements based on machine vision,the design and implementation of the basic functions of the analysis system are completed,and the goal of providing quantitative data information to medical staff is achieved.The quantified evaluation data information is calculated and obtained from the rehabilitation action video stream,so that the medical staff can more accurately judge the patient’s rehabilitation status,and confirm or adj ust the follow-up rehabilitation plan in time;Make a synchronous analysis and understanding to facilitate the follow-up rehabilitation training.The analysis system studied in this paper can be applied to the home environment.The method used does not rely on additional wearable sensors and depth cameras.The data-based measurement and analysis of the quality of rehabilitation actions can be achieved by relying on simple mobile phone cameras,which has a very good application.prospect.