摘要
基于视觉的手势识别是实现新一代人机交互的关键技术。本文提出了一种使用深度信息进行手指检测和手势识别的研究方案。利用微软Kinect传感器获取图像深度信息，通过阈值分割和K-均值聚类算法获取手部图像。手型外部凸包检测结合手部外部轮廓检测实现指尖的确定，根据每个指尖到掌心的矢量和手指间的相对位置关系实现手指识别，通过手指的分类实现对数字手势1～5的手势识别。实验结果表明，该研究方案能准确检测手指，手势识别率超过95％。
Vision-based gesture recognition is a key technique to achieve a new generation of human-computer interaction. This paper proposes a research program for fingertip detection and gesture recognition using depth information. The depth information of an image is captured using Microsoft access sensors, and threshold segmentation combined with the k-means clustering algorithm is used to obtain the hand image. The shape of the hand external convex hull detection combined with hand external contour detection is used to determine the fingertip. Vectors between each of the fingertips and the palm as well as the relative position of the relationship between the fingers are used for finger recognition. The finger gesture recognition of digital signal of 1-5 is conducted through the classification. The experimental results show that the research scheme can accurately detect finger, and hand gestures recognition rate is more than 95%
基金
国家基金资助（61171152）,浙江省自然科学基金项目资助（LY13F020044）
关键词
手势识别
指尖检测
人机交互
深度信息
Hand Gesture Recognition
Fingertip Detection
Human-computer Interaction
Depth Data
