Vision-Based Posture Detection for Rehabilitation Program
Sudhir Gaikwad, Shripad S. Bhatlawande, Atharva Dusane, Dyuti Bobby, Krushna Durole, Swati Shilaskar · Advances in engineering research/Advances in Engineering Research · 2023
Individuals with disabilities frequently struggle to do simple tasks.Recurrent workouts have been demonstrated to aid affected patients in rehabilitation.Physical rehabilitation therapy that can be self-managed provides a convenient solution for people with motor disabilities who may find it challenging to attend regular in-person therapy sessions.Analyzing body postures is instrumental in assisted living and health monitoring at home.Tracking body postures is a profound issue in computer vision.Monitoring the upper-limb posture of the body is the primary goal, while considering the complication of human pose, despite having no publicly available dataset.In this text, a self-data procurement system followed by real-time body posture recognition is implemented using LSTM.The posture classification accuracy is 93.75 percent.If current frames are incorrect, immediate results will be displayed.As a result, the user can instantly improve their posture if they complete their exercises inaccurately, by viewing the correctness of their performance in real-time.