Three Dimensional Hand Movement Tracking with Machine Learning and Stereo Camera

A novel methodology is suggested to track the hand movement in 3-dimensions, using a computer vision based machine learning model to identify various points on the hand and doing it repeatedly in each frame . The proposed method is based on Stereo camera setup to calculate the change in depth, in which, the hand is recognized at different points using machine learning in both cameras and is used to calculate the change in position with respect to a initial position. The Computer vision machine learning model is implemented on python. The displacement is recorded with respect to the starting position of the hand initialized, co-ordinates are printed on the camera window as well as exported to a CSV file. The various positions of the hand are observed and the displacement from initial recorded.