Abstract: Based on the configuration chacters of the finger vein image，this paper proposes an enhanced method for extracting finger vein features，the basic principle is to extract features by detecting concavity in gray image.Experiments proves that it is not only simpleness，easy to be achieved and adaptive，but also novelty by compareing traditional feature extraction algorithm.Effect of extracting finger vein features is very perfect，and extracting features is accurate，continuous and derivable meanwhile.

Key words: finger vein, feature extraction, feature image, de-noising

摘要： 根据手指静脉图像的结构特点，提出了一种手指静脉特征的提取算法，其基本思想是利用灰度图像中静脉灰度值所形成的谷形区域来实现特征的提取。详细讨论了该算法的基本原理和具体计算步骤。实验结果表明：该算法与传统的特征提取算法相比，方法简单，易于实现，具有自适应性，而且思想新颖，对手指静脉的提取得到了非常理想的效果，提取的特征不但准确而且连续光滑。

关键词: 手指静脉, 特征提取, 特征图像, 去噪