A Human Gait Classification Method Based on Adaboost Techniques Using Velocity Moments and Silhouette Shapes

Authors:
Chin-Shyurng Fahn;Ming-Jui Kuo;Min-Feng Hsieh
Affiliations:
Department of Computer Science and Information Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, Republic of China 10607;Department of Computer Science and Information Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, Republic of China 10607;Department of Computer Science and Information Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, Republic of China 10607
Venue:
IEA/AIE '09 Proceedings of the 22nd International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems: Next-Generation Applied Intelligence
Year:
2009

Citing 3
Cited 0

Statistical Gait Description via Temporal Moments

SSIAI '00 Proceedings of the 4th IEEE Southwest Symposium on Image Analysis and Interpretation
Automated Detection of Human for Visual Surveillance System

ICPR '96 Proceedings of the International Conference on Pattern Recognition (ICPR '96) Volume III-Volume 7276 - Volume 7276
Statistical motion model based on the change of feature relationships: human gait-based recognition

IEEE Transactions on Pattern Analysis and Machine Intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we propose a human gait classification method based on Adaboost techniques that identify three kinds of human gaits: walk, run, and limp. We divide a video sequence into several segments, each of which is regarded as a process unit. For each process unit, we collect both the velocity and shape information of a moving object. We first apply the Canny edge detector to enhancing the edges within a foreground image, followed by computing the distance and the angle difference between two edge pixels which are put into the accumulation table. The gait classification employs an Adaboost algorithm which is excellent in facilitating the speed of convergence during the training. The experimental result reveals that our method has good performance of classifying human gaits using both the velocity and shape information.