A Novel Preprocessing Algorithm of Knuckleprint

Knuckle print is a new biometric technology and contact less one is a convenient and cleaner manner. Extracting knuckle print from the original image is important to the improvement of identification rate. This paper introduces a new method of binarization and segmentation for contact less knuckle print authentication. Firstly, the RGB image is converted into HSV color space according to the clustering of hue, and then k-means algorithm is adopted to reduce the light impact during the binarization. Subsequently, a new corner detection approach is put forward in order to establish the reference coordinate. Finally, the Region of Interest (ROI) that is a rectangle including the whole single finger is obtained. The proposed algorithm has been tested on an image database using contact less acquisition in natural lighting. The experimental results demonstrate the proposed algorithm actually improved the effect of segmentation.