摘要： 鉴于人体的手指指节折痕具有稳定性且因人而异的特点，该文提出一种基于投影和小波分析的手指指节折痕识别的新算法。对手指图像进行定位，经分割、归一化后得到了用于身份鉴别的手指子图。将手指子图向水平轴投影，得到含有折痕信息的一维信号，利用小波的多分辨率特点提取折痕信息进而形成代表折痕的特征矢量。在手指图像数据库中，利用最近邻分类器进行了算法验证，等错误率约为1.5%，实验结果表明了算法的有效性。

关键词: 手指指节折痕, 生物特征识别, 小波分析, 特征提取

Abstract: Because the finger creases are thought to be stable and different for each individual, this paper proposes a new finger crease recognition method based on projection and wavelet. After image localization and normalization, the subimage for each finger image is obtained. The subimage is a rectangular window on inner side of the finger, which includes the first and the second joint line in the finger image. The subimage is projected onto horizontal axis to construct one-dimensional signal. Wavelet analysis is applied to extract finger crease features from one-dimensional signal to form the feature vector. The proposed method has been tested on a finger database by using the nearest neighbor classifier. The experimental results show the effectiveness of the proposed method in terms of the Equal Error Rate (EER) (≈1.5 percent).

Key words: finger crease pattern, biometrics, wavelet analysis, feature extraction

中图分类号: 

• TP391.4