Hand Vein Recognition Based on Improved Template Matching

Abstract

An improved template-construction method for hand vein recognition is presented. The accomplishment of the new algorithm involves several steps. First, we design the power-controlled multispectral vein acquisition handset to establish high-quality hand vein database. Second, we propose an improved valley-shaped enhancement operator with NiBlack algorithm to realize vein segmentation. Third, we get a basic-shaped template after dilation and erosion operation in the four neighborhoods firstly, and then we can get the final matching template combining all the shapes after the process of filling the connected region whose threshold is smaller than the one we set according to the experiment effect. Finally we design the hand vein recognition experiments using the designed template and traditional template respectively. The recognition rate of the improved template is 96.6% under the 1:1 mode and 95.75% under the 1:N mode while the recognition rate of the traditional template is 89.2 under the 1:1 mode and 88.2 under the 1:N mode.

Keywords

• Vein recognition
• Template matching
• Power-controlled
• Multispectral
• Vein segmentation
• Enhancement operator