This work deals with 3D hand tracking in cluttered background which is an important task in human computer interaction and non intrusive marketing behavior. In this paper a particle filter framework is proposed to integrate gradient distributions and image observations in order to estimate the 3D position of hand from monocular image sequences. Extensive experiments have bee carried out to demonstrate the efficiency and the robustness of our approach.
Published in:
Intelligent and Advanced Systems (ICIAS), 2012 4th International Conference on
(Volume:1
)
Date of Conference: 12-14 June 2012