Monocular 3D human body reconstruction towards depth augmentation of television sequences

This paper addresses the reconstruction of 3 0 human body models fram Z D video sequences. Considering that the input frames are already segmented, the proposed technique consists of three stages. These stages are independently applied over each segmentedframe. Firstly, a skeleton of a humanfigure obtainedfrom the segmented image is extracted by means of afast algorithm based on a Voronoi diagram of the boundary paints. Afterwards, the skeleton is labelled according to the human body parts (e.g. head, upper arm, lower arm, torso, etc). Secondlfi an initial 3 0 model posture is estimatedfrom the labelled skeleton. Finally, an iterative closest point (ICP) implementation is used to refine the initial model posture by maximizing the similarity between the projected 3 0 model and the segmented image. Experimental results with video sequences are presented.

• Type: Conference
• Year: 2003
• Where: ICIP
• Authors: Angel Domingo Sappa, Niki Aifanti, Sotiris Malassiotis, Michael G. Strintzis
• Keywords: Human Body, ICIP 2003, Image Processing, Segmented Image, Video Sequences