A Semi-Supervised Approach to Human Pose Estimation

Human Pose Estimation

Human pose estimation deals with the problem of estimating a 2D or 3D skeleton of the human body, given some input data. The data usually consists of an image or a video of one or more people. Being a basis for many applications (such as physiotherapy, sports science, animation and game industry …)  and research fields (like activity and action recognition, video surveillance, …) this field has gained much attention in the last decade. Major work in this area can be divided into generative and discriminative approaches.

This Project:

DML’s Human pose estimation research group has focused on discriminative (Learning based) approaches to single camera, full body human pose estimation. We try to develop new learning algorithms or adapt the existing ones for human pose estimation. You may find some important papers in this field in the papers section. Some links to some of the major databases for human pose estimation is given in the links area.

People involved: Hamid R. Rabiee, Nima Pourdamghani