Fingerprint skeletonization extraction based on improved principal graph algorithm
利用改进的主图形算法提取指纹骨架

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.