基于二阶段小波多分辨率分析的掌纹分割算法

针对复杂掌纹纹线难以分割、有效性低的问题,提出一种基于二阶段小波多分辨率分析的掌纹分割算法.该方法首先利用小波多分辨率分析高频子图的候选子区域,对得到的相似掌纹纹线集合进行合并；接着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得到掌纹主要纹理特征.实验结果表明,该方法不仅能有效地剔除复杂掌纹的噪声,而且能准确提取掌纹特征,从而达到准确识别的目的.

PALMPRINT SEGMENTATION ALGORITHM BASED ON TWO PHASE WAVELET MULTI-RESOLUTION ANALYSIS

In order to solve the problems that the lines of palmprint are hard to be segmented and the low segmentation effectiveness,this paper proposes a palmprint segmentation algorithm based on two phase wavelet multi-resolution analysis.The algorithm first analyses the candidate sub-regions of high frequency sub-image using wavelet multi-resolution and merges the derived similar palmprint lines sets,and then seeks the intersection of the merged similar regions set and the binary set to obtain the fused image,finally attains main texture features of palmprint using regional growth method and morphological denoising.Experimental results show that this method can remove the noises in complex palmprint effectively,and can also extract the palmprint feature accurately,so that achieves the purpose of accurate identification.