低对比度手指静脉图像的分割

摘要

In order to overcome the imprecise of low contract finger vein image by using segmentation algorithm, the author adopts local dynamical threshold algorithm; designs new variable weight decision function according to relativity of gray value of vein image in spatial information, and solves the fake vein problem after segmentation. Besides, the author puts forward new criteria, which can solve the pixel redundancy problem effectively, aiming at redundancy problem in vein intersection caused by general thinning algorithm. Experiments indicate that the proposed methods can segment the low contract finger vein image effectively.%为了克服目前分割算法中低对比度手指静脉图像分割不精确的问题,采用局部动态阈值算法,依据静脉图像中灰度值在空间信息的相关性,设计了新的变权重判定函数,解决分割后产生的伪静脉问题.同时针对一般细化算法在静脉交叉点产生的冗余问题,提出了新的判断准则,有效地解决了像素冗余问题.实验表明,这些方法能有效地分割低对比度手指静脉图像.