%0 Conference Proceedings
%A Wang, Ruifang,
%A Ramos, Daniel,
%A Fierrez, Julian,
%A Krish, Ram P.,
%@affiliation Universidad Autonoma de Madrid
%@affiliation Universidad Autonoma de Madrid
%@affiliation Universidad Autonoma de Madrid
%@affiliation Universidad Autonoma de Madrid
%T Towards Regional Fusion for High-Resolution Palmprint Recognition
%B Conference on Graphics, Patterns and Images, 26 (SIBGRAPI)
%D 2013
%E Boyer, Kim,
%E Hirata, Nina,
%E Nedel, Luciana,
%E Silva, Claudio,
%S Proceedings
%8 Aug. 5-8, 2013
%J Los Alamitos
%I IEEE Computer Society
%C Arequipa, Peru
%K High resolution palmprints, regional fusion.
%X The existing high resolution palmprint matching algorithms essentially follow the minutiae-based fingerprint matching strategy and focus on full-to-full/partial-to-full palmprint comparison. These algorithms would face problems when they are applied to forensic palmprint recognition where latent marks have much smaller area than full palmprints. Therefore, towards forensic scenarios, we propose a novel matching strategy based on regional fusion for high resolution palmprint recognition using regions segmented by major creases features. The matching strategy includes two stages: 1) region-to-region palmprint comparison; 2) regional fusion at score level. We first studied regional discriminability of a high resolution palmprint under the concept of three regions, i.e., interdigital, hypothenar and thenar, which is the most significant difference between palmprits and fingerprints. Then we implemented regional fusion based on logistic regression at score level using region-to-region comparison scores obtained by a commercial SDK, MegaMatcher 4.0. Significant improvement of recognition accuracy is achieved by regional fusion on a public high resolution palmprint database THUPALMLAB. The EER of logistic regression based regional fusion is 0.25%, while the EER of full-to-full palmprint comparison is 1%.
%@language en
%3 Camera_Ready_Towards Regional Fusion for High Resolution Palmprint Recognition.pdf