An efficient approach for classification of palmprint images using heart line features

Palm print is one of the important hand related biometrics characteristics with high user acceptance. A new classification approach using heart line feature of palm print is proposed in this paper. The hand image captured from digital camera is preprocessed to find palm print Region of interest (ROI). Gabor filters are applied on palm print ROI and line detection operation is proposed to extract heart line features. Then the palm print images are classified based on the shape feature of the heart line. Proposed classification approach categories palm print images into four categories. Testing of the proposed approach on the database collected at our institute shows that the identification accuracy of 94% is obtained. Palm print features are extracted by applying local binary descriptor on palm print ROI and matched with euclidean distance in the most potential palm print category and if necessary matching process continues orderly to less potential categories. Identification experiment results show that the proposed approach can reduce the number of templates considered for matching from 100% as in the case of conventional approaches to a range of 30% to 60% while maintaining the same accuracy as the that of conventional approaches.