HAND GESTURE RECOGNITION BASED ON FREE-FORM CONTOURS AND PROBABILISTIC INFERENCE.

A computer vision system is described that captures color image sequences, detects and recognizes static hand poses (i.e., "letters:) and interprets pose sequences in terms of gestures (i.e., "words"). The hand object is detected with a double-active contour-based method. A tracking of the hand pose in a short sequence allows detecting "modified poses", like diacritic letters in national alphabets. The static hand pose set corresponds to hand signs of a thumb alphabet. Finally, by tracking hand poses in a longer image sequence, the pose sequence is interpreted in terms of gestures. Dynamic Bayesian models and their inference methods (particle filter and Viterbi search) are applied at this stage, allowing a bi-driven control of the entire system. [ABSTRACT FROM AUTHOR]