A rough-set based binarization technique for fingerprint images

Fingerprint is considered as the most robust biometric in the sense that it could even be obtained without willing participation of the subjects (uncontrolled situation). Fingerprints are unique in the sense of location and direction of minutiae points present. A set of minutiae points thus best characterize the fingerprints. However extracting minutiae points from fingerprint under uncontrolled situation is a challenge. A very robust binarization process is of high demand to get correct set of minutiae points. In this paper, a rough-set based approach for binarization of fingerprint image is presented. Maximization of rough entropy and minimization of roughness of the image lead to an optimum threshold for binarization. The result of the proposed method is compared with the traditional Otsu's thresholding method for binarization.