A Bayesian approach to human activity recognition

This paper presents a methodology for automatically identifying human action. We use a new approach to human activity recognition that incorporates a Bayesian framework. By tracking the movement of the head of the subject over consecutive frames of monocular grayscale image sequences, we recognize actions in the frontal or lateral view. Input sequences captured from a CCD camera are matched against stored models of actions. The action that is found to be closest to the input sequence is identified. In the present implementation, these actions include sitting down, standing up, bending down, getting up, hugging, squatting, rising from a squatting position, bending sideways, falling backward and walking. This methodology finds application in environments where constant monitoring of human activity is required, such as in department stores and airports.