摘要
手语识别的研究起步于20世纪90年代,国内外知名高校、研究机构和学者在手语识别技术上都投入很大研究力量。采用图像处理方法,从视频图像中,通过提取肤色,图像形态运算,边缘提取,对图像进行预处理,然后对每一帧图像中的手势轮廓数据计算Hu矩,来采集手势特征,最后将动态时间规整(DTW)应用到图像识别中,对0-9十个数字进行检测,实验中,取得较好的识别效果。
Sign language recognition research started in 1990s, domestic and foreign well-known universities, research institutions and scholars have put a lot of effort into this. Uses the image processing method, find the contour of gesture, and calculates Hu moments as the feature of each gesture, applies the dynamic time warping （DTW） method to the detection, to recognize the gesture of number 0 to 9. The experiment results show that the algorithm has a good performance.
关键词
手势
DTW
边缘检测
HU矩
Gesture Recognition
DTW
Edge Detection
Hu Moments
