Upper-body contour xtraction and tracking using face and body shape variance information

We propose a fitting method using a model that integrates face and bod□ shape variance information for upperbod□ contour extraction and tracking. Accurate bod□-contour extraction is necessar□ for various applications, such as pose estimation, gesture recognition, and so on. In this stud□, we regard it as the shape model fitting problem. A model including shape variance information can fit to the contour robustl□ even in the nois□ case. AAMs are one of these models and can fit to a face successfull□ It needs appearance information for effective fitting, but it can not be used in our case because appearance of upper-bod□ easil□ changes b□ clothes. Instead of intensit image, proposed method uses edge image as appearance information. However, discrimination between a true contour edge of upperbod□ and other edges is difficult. To solve this problem, we integrate shape models of upper bod□ and face. It is expected that this integrated model is more robust to edges in clutter background and various locations of the bod□ than a bod□ shape model using onl□ bod□ shape information. We conduct experiments and confirm improvement in accurac □b□ integration of face and bod□ variance information.