DEEP POSE AND HUMAN POSE ESTIMATION VIA NEWRAL NETWORK

Abstract

Deep Neural Networks are used to assess an individual's posture, and we present a technique for doing so. Deep Neural Networks are used to assess an individual's posture (DNNs). Regarding the subject's body joints, it is argued that the pose estimation problem may be conceived of as a DNN-based regression problem in terms of the subject's posture. In this paper, it is proven how to design a cascade of such DNN predictors, which leads in high precision position predictions for the target location. Pose reasoning can be completed in its entirety with the help of this technique, which has a clear yet strong formulation that takes advantage of the most current breakthroughs in deep learning technology to do this. Using four academic benchmarks with diverse real-world photographs, we present a complete empirical analysis that reveals state-of-the art or higher performance on four academic benchmarks, as proven by the findings of four academic benchmarks.