動画像からの手振り動作識別のための動作部分領域とその手形状抽出

抄録

Recently, there are many reports to recognize human gesture by using image processing as new man-machine interface with construction of virtual world or recognition of hand sign. In most particular, auto-recognition of hand gesture having intuitive and wide power of expression is very important. Extremely precious and stable extraction of figure of moving part and its hand shape is one of important points, if it could be done, it is possible to extract necessary and important information for hand gesture discrimination. In this report, a method which extracts figure of moving part and its hand shape from sequence images by making use of only image processing is proposed and availability of the method is confirmed by experimentation using actual images.