Abstract: Pointing gesture recognition can be taken as one of optimal human computer interaction models.Efficient extraction of pointing fingers is a key and difficult problem.A new，simple and fast fingertip extraction method is developed.The foreground target is segmented based on background subtraction at first.Secondly，the skin color model is used to extract face and hand regions.Based on position relationship between the face and hand and the fact that fingertip locates at the contour of hand region，pointing fingertip is chosen from those points which are intersection of boundary rectangle and hand contour.Experimental results show that the developed method can extract pointing fingertips accurately and efficiently.

Key words: pointing gesture recognition, fingertip extraction, background, segmentation, real time

摘要： 指势行为识别作为一种理想的人机交互模式，而指势手指有效分割与指尖提取是关键。采用基于背景差分方法提取前景运动目标并消除背景影响，运用肤色分割方法，提取运动目标脸部、手部区域。在确定指势手与人脸位置关系的基础上，基于指势行为中指尖位于指势手的边沿轮廓，从指势手的外界矩形框与轮廓的交接点中定位指尖点。经实验证明该算法快速有效、提取精度高，且有一定的鲁棒性。

关键词: 指势识别, 指尖提取, 背景, 分割, 实时