A Local Entropy Based Palmprint Image Enhancement Algorithm

This paper proposes a promising palmprint image enhancement algorithm. Under the constraint that keeps original characteristic of palmprint, the method improves contrast of images, in order to extract features from palmprint by Scale Invariance Feature Transformation (SIFT) descriptor. However, the existing image enhancement algorithms can hardly yield SIFT feature from low resolution palmprint images, e.g., dpi of these images is less than 150. Our scheme uses local entropy to reassign the enhancement coefficients of traditional Unsharp Mask (UM) algorithm, for an enhancement of palmprint effectively. The algorithm solves the problem of obtaining SIFT keypoints from enhanced palmprint images successfully, which is failure by traditional UM based schemes or histogram equalization.