Real Time Pose Estimation

• Improved the efficiency of physiotherapists by implementing real-time solution from the live camera feed.

• Solution evolved from the basic use-case of sports-person suffering from injuries due to inappropriate angle at the time of the sports.

• Our solution focused on the real-time video analysis and generating insights based on the angles of different body parts by predicting the pose using Residual Network.

• Two-point Cosine Algorithm helps us to find out the angle from the pose which is used by the physiotherapist to identify the appropriate pose and increase the efficiency of sportsman.

• Using Linear Regression, we have also identified the normal angle physiotherapists can suggest for better performance.