Two-stage strategy for palmprint identification

A novel approach for palmprint identification based on a two-stage strategy is proposed. First a new method for palmprint alignment and segmentation is presented. The largest center area of the palmprint is extracted as the region of interest (ROI), and geometric features are extracted as the matching features of rough matching Then a new approach of line feature extraction and representation is proposed for exact matching Palmprint identification classifier from rough matching to exact matching is designed, in which the rough matching algorithm is employed to extract the subset of palmprint database classes and the exact matching is achieved by line feature. Experiments were performed on a database of 250 images from 50 individuals. The identification accuracy is up to 94.67%, proving the effectiveness of the proposed algorithm.