An improved square-based palmprint segmentation method

Abstract

Palmprint segmentation is an important task for palmprint recognition. By the segmentation operation, an expected area of each palmprint can be extracted for the future feature extraction and matching. In the paper, an improved square-based segmentation method for palmprint alignment and extraction is presented. The approach successfully extracts three key points in a local area. The outside boundary of palm is approximated as a reference line to align palmprint images and the reference line is decided by these key points. In the aligned palmprint image, a coordinate system is set up and a central part of palmprint image is extracted. In addition, the method can automatically adjust the size of the extracted area according to the width of the palm. The experimental results demonstrate the effectiveness and accuracy of segmentation of the proposed approach.