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 Network s
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 formula t ion
that takes advantage of the most current breakthroughs in d eep
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.