手形识别中的手形提取方法

手形识别在生物特征识圳领域中占有重要的地位，目前最常用的手形匹配方法是点匹配的方法，手形提取的好坏直接影响着识别的准确性，这就要求于形的轮廓提取必须准确，否则会引入大量的伪特征点从而带来错误率的激增。在手形采集过程中，周围环境的光照和噪声对于形的准确提取会造成很大的影响。该文提出了一种光照补偿和由基于分类搜索的局部域值手形提取方法解决以上问题，该方法能有效地克服传统方法难以提取手形边缘信息的现象，实验表明，该方法能准确有效地对手形讲行提取，为识别提供了保障。

Method of Hand Shape Extrcting in Hand Shape Recognition

Hand shape recognition is an important measure in biometrics. Currently, the most popular method of hand shape matching is point pattern matching, and the way of hand shape extracting determines the accuracy of hand shape recognition. The illumination and the noise will make the extracting success rate drop quickly. In this paper, we present an idea of illumination compensation and local-threshold segment based on class searching to extract hand shape features. This algorithm can get over the issue that it is not easy to extract the edge in traditional ways. The experiment data proves that the idea can extract the hand shape features accurately and be used in the face extracting and palm locating effectively.