Newfangled 3d human pose estimation using MediaPipe with foreground object detection
R. Anand, Malini Mahendiran, Shaik Shahrukh Ahmed · AIP conference proceedings · 2022
Human pose assessment from video plays an important role in various applications such as measuring physical exercises, sign language recognition and full body gesture control. For example, it can be the basis for yoga, dance, and exercise applications. It can also enable the overview of digital content and information at the top of the physics world in augmented reality. Here we propose, Newfangled 3d Human pose estimation using MediaPipe with foreground object detection (HPEM), model uses MediaPipe library. MediaPipe pose is applied for faster pose estimation of humans, prior in which the humans and other objects are detected and classified in foreground object detection. This paper conducted various experiments, and the results proved superior and the success of this algorithm.