利用改进的主图形算法提取指纹骨架

Fingerprint skeletonization extraction based on improved principal graph algorithm

在传统的指纹识别系统中,一般使用指纹细化图对指纹信息进行描述。尝试使用主曲线代替指纹细化图,并对主曲线的主图形算法进行了较大改进,使其能够得到更好的指纹骨架。实验结果表明,主曲线骨架比细化图有着光滑性强、准确性高、抗噪音和信息量大的优点;而改进的主图形算法在指纹骨架提取的效率和效果上比原算法也都有一定提高。

The principal curve was used to describe and reflect the structure of fingerprint instead of the thinned fingerprint image as in the traditional fingerprint recognition system. And the principal graph algorithm was developed to get a better skeletonization of fingerprint. The experiment results show that the improved principal graph algorithm is better than the original algorithm in efficiency and quality. Compared with thinned fingerprint image, the new method is determined to be smoother, more accurate, and more anti-noisy, and has more information quantity.