Abstract
In this paper, we propose the method of human gait estimation using a wearable camera. The proposed method has an advantage in achieving the estimation regardless of surroundings. Human gait is estimated based on Extended Kalman Filter by using some human walking models which are defined as example motion clips, and an observed image which is captured by the camera attached on human leg. The experiment shows that this method allowed the accurate estimation even when an user walked at different velocity.
