基于融合边缘算子和最大熵的低质量指纹分割研究  

Based on the Fusion Edge Operator and Maximum Entropy Low Quality Fingerprint Segmentation Research

作　　者：张升斌[1] 李吉成[1] 余怀田[1] 

机构地区：[1]国防科学技术大学,湖南长沙410073

出　　处：《数字技术与应用》2013年第1期63-64,67,共2页Digital Technology & Application

摘　　要：针对指纹低质量指纹图像分割的特征,并总结了常见指纹分割算法存在的缺点,引出采用边缘检测的方法进行低质量指纹分割算法的设计研究。本文考虑由于单纯采用Sobel-Snake算子仅能起到平滑和抑制噪声的作用,并不能实现指纹图像和背景的分离,故此在Sobel-Snake算子检测之前采用二维最大熵的方法对图像进行二维最大熵技术,然后通过Sobel-Snake算子获取边缘图像,并根据给定阈值向量实现了低质量指纹分割算法。最后给出实验效果和分割图像的评价指标。For the characteristic fingerprint of low-quality fingerprint image segmentation, and summed up the shortcomings of the common fingerprint segmentation algorithm !eads to the edge detection method for the study of low-quality fingerprint segmentation algorithm. In this paper, we consider the simple Sobel-Snake operator can only play the two-dimensional maximum entropy smoothing and noise suppression effect, and cannot achieve the separation of the fingerprint image and the background, and therefore before the Sobel-Snake operator to detect two-dimensional maximum entropy method of image technology, and then obtain the edge images by Sobel-Snake operator, and low-quality fingerprint segmentation algorithm to achieve a given threshold vector. Finally, the experimental results and evaluation of image segmentation.

关 键 词：低质量指纹图像 Sobel—Snake算子 二维最大熵 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]