摘要
二值化是数字指纹自动识别中不可或缺的重要环节,阈值的每一个像素的变化都会影响二值化处理效果,并且会直接影响到特征数量的保持度。提出了一种新的相位移动模糊差分的指纹图像二值化处理方法,使得前景和背景的分割点趋向直方图的0值,克服了敏感阈值问题并保持了原有的特征点。
Binarization is one important step in automatic digital fingerprint identification. Selection of threshold will greatly influence the effect of binarization, as well as the number of valid features. In this paper, we proposed a new phase remove smoothing and difference method for fingerprint image binarization, which moves the threshold of background and foreground toward 0 point. Experiments show that this method can reduce the sensitivity of threshold selection, and at the same time can keep the valid fingerprint features after binarization.
出处
《计算机应用》
CSCD
北大核心
2007年第1期169-171,共3页
journal of Computer Applications
基金
国家自然科学基金资助项目(60175001)
关键词
指纹
相位移动
模糊
差分
二值化
fingerprint
phase remove
smoothing
difference
binarization
