中文摘要：

为了实现视频手势的实时分割与定位，提出一种基于肤色采样点筛选的视频手势分割算法对视频中的运动手势进行实时分割与定位。首先，对视频的每一帧通过图像预处理获取肤色采样点；其次，融合运动信息对肤色采样点进行初步筛选；然后，通过区域生长法再次筛选肤色采样点提取特征点；最后，通过特征点实现视频手势的分割与定位。实验结果表明，该方法具有良好的手势分割效果和较高的定位准确度。

英文摘要：

In order to realize real-time segmentation of video gestures and location, in this paper a method of the segmentation and location of motion gestures in real-time video is presented based on screening skin color sample points.Firstly, in each frame of videos, obtaining color sample points by Image preprocessing.Secondly,the color sample points are screened for the first time with the movement information.Then, feature points are acquired by screening the color sample points for the second time by using region growing method. Last but not least, realizing segmentation of the video gesture and location with the feature points. It is proved that the method has good segmentation effect and higher positioning accuracy.