Rehabilitation robot and amusement robot look forward to being active at Medical and welfare field accompanied by a society composed largely of elderly people. On the other hand, some situations of human gesture express human a mental state at times. In this context, we examined the tracking system of human gesture for the purpose to realize for amusement robot by analyzing the image motion from a moving image sequence. This tracking system is composed that generalized gradient method for the Optical Flow and stochastic motion modeling according to statistical inference by Bayes'theorem. We demonstarated the tracking system by experiments using real image sequence.
展开▼
