A real-time gesture tracking and recognition system based on particle filtering and Ada-Boosting techniques

Authors:
Chin-Shyurng Fahn;Chih-Wei Huang;Hung-Kuang Chen
Affiliations:
Department of Computer Science and Information Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, R.O.C.;Department of Computer Science and Information Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, R.O.C.;Department of Electronic Engineering, National Chin-Yi University of Technology, Taichung, Taiwan, R.O.C.
Venue:
UAHCI'07 Proceedings of the 4th international conference on Universal access in human-computer interaction: ambient interaction
Year:
2007

Citing 5
Cited 0

Linear-time connected-component labeling based on sequential local operations

Computer Vision and Image Understanding
Multi-Modal System for Locating Heads and Faces

FG '96 Proceedings of the 2nd International Conference on Automatic Face and Gesture Recognition (FG '96)
Color-Based Hands Tracking System for Sign Language Recognition

FG '98 Proceedings of the 3rd. International Conference on Face & Gesture Recognition
Hand gesture recognition using depth data

FGR' 04 Proceedings of the Sixth IEEE international conference on Automatic face and gesture recognition
Real time hand tracking by combining particle filtering and mean shift

FGR' 04 Proceedings of the Sixth IEEE international conference on Automatic face and gesture recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

A real-time gesture tracking and recognition system based on particle filtering and Ada-Boosting techniques is presented in this paper. The particle filter, which is a flexible simulation-based method and suitable for non-linear tracking problems, is adopted to achieve hand tracking robustly. In order to avoid the influence of the other exposed skin parts of a human body and skin-colored objects in the background, our system further applies the motion information as a feature of the hand in addition to the skin color information. Compared with the conventional particle filters, our method leads to more efficient sampling and requires fewer particles. It results in lowering computational cost and saving much time for gesture recognition later. The gesture recognition uses the features derived from the wavelet transform, and employs an Ada-Boost algorithm which is excellent in facilitating the speed of convergence during the training. Hence, it is conducive to update new information and expand new gesture archives. The experimental results reveal our system is fast, accurate, and robust in hand tracking. Moreover, it has good performance in gesture recognition under complicated environments.