Orientation space code and multi-feature two-phase sparse representation for palmprint recognition

Orientation based coding method is one of the most important approaches in palmprint recognition, which achieves impressive performance by extracting one or more dominant orientation features, and calculating the distance between features of two palmprints. However, simply using the orientation features may be vulnerable to noisy and rotation. In this paper, we proposed a novel orientation-space code scheme to represent the orientation space feature of palmprint and designed a novel multi-feature two-phase sparse representation (MTPSR) scheme for feature matching. Extensive experiments on three benchmark palmprint databases are conducted to demonstrate the high effectiveness of the proposed method