Prediction of Basketball Free Throw Shooting by OpenPose

Abstract

OpenPose, which is developed by Carnegie Mellon University (CMU) presented in CVPR 2017, takes in real-time motion images via a simple web camera and is capable of recognizing skeletons of multiple persons in these images. It also generates recognized skeleton point coordinates to files. OpenPose is featured by CMU’s original top-down method for real-time recognition and it is open online especially for research purposes. Thus we aimed to build a posture analysis model using OpenPose skeletal recognition data and verifying the practicality of OpenPose by verifying the accuracy of the model. As a posture analysis model, we adopted a logistic regression model that predicts the shooting probability of the basketball free throw with skeleton posture data as explanatory variables and the fact whether the ball enters the basket or not as a binary target variable. As the result, sufficiently significant prediction accuracy was obtained. Therefore, posture analysis using OpenPose has been verified to be practical with our model. We consider that with many skeleton data which are easily provided by a simple web camera, OpenPose makes statistical diagnostic approach possible. We also consider it could lower costs (in both financial and time-wise) of such an analysis which has previously required more equipments and more time for preparation regarding motion capture analysis systems.