一种自适应块大小的指纹图像分割算法

摘要

指纹图像分割是影响指纹识别系统性能的关键步骤之一.为此,研究指纹图像频谱图和传统的基于块的指纹图像分割方法,由频谱图上能量分布情况得到平均指纹纹线距离,结合生物统计学原理把指纹图像分成互不重叠的矩形块,块大小根据平均指纹纹线距离在算法执行过程中自动确定.实验结果表明,该算法对不同分辨率及不同质量指纹图像的适应性都高于传统算法.%Fingerprint segmentation is one of the key steps of fingerprint identification system, which are relation to its performance. It researches the spectrogram and traditional segmentation method based on block of fingerprint images, and calculates the average fingerprint ridge distance,according to spectral energy distribution of the map, then proposes to divide the image into non-overiapping rectangular block according to physiological characteristics of the human body. The block size is automatically determined in the algorithm implementation process, according to the characteristics of frequency spectrum. Experimental results show this algorithm is adaptive to be significantly higher than the traditional ones, for images of different resolutions and quality.