Fingerprint identification based on semi-supervised FSS

Fingerprint images captured in real world applications always include some variations, called intra-class variations, due to various uncontrolled conditions like scratching, aging, moisting, drying, etc. It is important for current fingerprint identification systems to adaptively deal with these variations. In this paper, we propose a semi-supervised FSS based fingerprint identification method. We use unlabeled samples to train FSS Center for each finger in a semi-supervised setting, which significantly improves the robustness of the FSS based method. We evaluate our method on the DIEE Fingerprint database. The experimental results show favorable performance of our method as compared to state-of-the-art.