摘要
由于复杂掌纹纹线存在难以分割和有效性较低的问题,该文采用小波多分辨率分析方法构建掌纹分割算法,算法首先利用小波多分辨率对高频子图的候选子区域进行分析,将获得的相似掌纹纹线集合加以合并,并对合并相似区域的集合和二值化集合求交集得到融合图像,最后利用区域生长法和形态学去噪获得掌纹主要纹理特征.
In order to solve the difficulty and low effectiveness of palm print segmentation,this paper proposes a new segmentation algorithm of palm based on wavelet multi-resolution analysis.Based on wavelet multi-resolution analysis of palm image high frequency candidates for the subregion,this method gets similar palm print set,and then incorporates dynamic threshold segmentation set and similar palm print set to locate palm print region.Finally the palm characteristic could get from denoise process such as regional growth and morphological thinning method.
出处
《湘潭大学自然科学学报》
CAS
CSCD
北大核心
2012年第1期117-121,共5页
Natural Science Journal of Xiangtan University
关键词
复杂掌纹
小波多分辨率分析
分割算法
区域生长法
complex palm
multi-resolution analysis
segmentation algorithm
regional growth
