基于小波理论对掌纹主线提取和修复

掌纹识别作为生物特征技术的一个重要分支，以其信息丰富、采集简便和稳定性强而作为疾病早期诊断和身份识别的重要依据，而其中主线的提取一直是识别的关键。本文讨论了一种提取主线信息的新方法，和传统方法不同的是，它利用普通的光学扫描仪方便地采取图源，之后对预处理和正规化后的图像利用Symlet小波变换的理论，提取掌纹主线的4个方向分量进行合并，并且利用ASF（alternating sequential filter）对结果进行形态学处理。结合回归分析和图像融合的方法，有效地消除了结果中出现的断裂区域，成功地从掌线中分离出了主线。将该方法与以前提取主线的方法进行了效果对比，同时，对于不同类别的掌纹都进行了主线提取验证，其结果说明了这种方法的强鲁棒性。该方法提取的主线信息为后期的临床诊断、皮纹分类，以及编码识别提供了有效和准确的数据。

Principle Line Extraction and Restoration Based on Wavelet Theory

As a vital branch in the study of biometrics-based technology, identification and verification by palm print has been a striking evidence for prior disease diagnosis and personal recognition, given its remarkable advantages like simplicity and stablility, etc. Especially, the extraction process of principle-line feature plays a key role. This paper presents a new approach to extract this novel characteristic. Unlike other traditional methods, its step is inherently simple and convenient using regular scanner. After the pre-process and alignment, we extract four spatial directional template and reach high convergence by adopting Symlet wavelets transformation method, and a series of morphological operations derived from ASF are utilized. Finally we use regression analysis and image fusion to eliminate divergence and disconnectedness in our result region, and successfully extract principle-line from numerous palm-lines. The experimental results with a large collection of different images showed its advantages compared with former work, and also illustrated its strong robustness, and provide effective and accurate statistics to clinical diagnosis, classification and encoding work at a later stage.