An Efficient Method for Human Behavior Identification

This paper presents a recognition method for human behavior identification based on motion history image theory. The motion history image has the advantage that it can record the motions of object over time. It can save complete motions history of object and have less computation for reckoning. The action features used in our system are motion gradient magnitude histogram and global/local motion orientation obtained from the motion history image. An error back-propagation neural network is used to identify the human behavior. The experimental results prove the feasibility and usefulness of the proposed method.

• Type: Conference
• Year: 2007
• Where: MVA
• Authors: Fang-Hsuan Cheng, Fu-Tai Chen
• Keywords: Computer Vision, Human Behavior, Motion Gradient Magnitude, Motion History Image, MVA 2007