There has been a lot of interest in constructing cognitive vision systems capable of detecting and identifying actions and activities. HMMs are perhaps the most successful framework in perceptual computing for modeling and classifying dynamic behaviors, popular because they offer dynamic time warping, a training algorithm, and a clear Bayesian semantics. This paper was motivated by the need of an efficient, robust, and real time HMM-based action recognizer, realistic in the sense that practical constraints such as the training set size are taken into account. Generalizing previous works, we formulate the main issues any recognizer should address, present a generic strategy to design such a recognizer, and formalize the conditions and the limitations on the features, as well as some important building steps and issues.
展开▼
