摘要
研究指纹图像,提取清晰图像处理过程,采集到的原始指纹图像,进行指纹特征提取。由于提取指纹是有方向性的,所谓分割算法就是将可用的指纹区域从背景噪声中提取出来。再采用方差法和方向图法处理基础上,提出了一种新的基于指纹方向流模式特性的SMDF分割方法。方法先将采集到的图像进行归一化,再对每个指纹图像块分别求取平行于块方向和垂直于块方向的灰度变化,选择作为前景块还是背景块的判断依据。实验研究表明,可以较为准确、完整地分割出指纹前景区,同时保留通过后续处理可以恢复的模糊区,还可以节省预处理的时间,也保证了精度,具有较强的可行性。
This paper proposes a new fingerprint segmentation algorithm based on the direction-flow after a profound study of the methods,which are based on gray scale and direction.Normalize the collected image,and then try to obtain the gray scale respectively on the parallel direction and perpendicular direction of the image grid.This is the judgment to distinguish foreground and background.Simulating result shows that the new algorithm can segment out fingerprint foreground precisely and completely.It can save a lot of pretreatment time and guarantee precision as well.
出处
《计算机仿真》
CSCD
北大核心
2010年第12期258-261,274,共5页
Computer Simulation
关键词
指纹分割
灰度差
方向流
规格化
Fingerprint segmentation
Gray scale
Direction flow
Normalize
