Recovering Human 3D Model from Monocular 2d Imaged for Detecting Postures Deformilities

Abstract

Human 3D pose estimation is one of the most talked-about or the hottest in recent years and even in the future. Most of the present algorithm can only get us the 3D skeletons from the pose which we have detected. There can be cases wherein because of lack of knowledge and data which we have obtained from the human shape, angles and textures, the accuracy of many various algorithms is not enough for high development level artificial intelligence and image processing tasks. In our paper, we purposed an algorithm and develop a project which will recover a 3D model of the human skeleton from the images which are captured by the camera to detect postural deformities and sitting postures. The algorithm proposed is very sturdy, dynamic and can help us for implementation of high development level image processing and computer vision tasks which also including shapes and textures. It also briefs us about many important techniques and important information from the implementation of the algorithm.