Logical Conjunction of Triple-Perpendicular-Directional Translation Residual for Contactless Palmprint Preprocessing

Contactless palmprint recognition systems alleviate the concerns on personal hygiene, acquisition flexibility, etc. Unfortunately, the preprocessing of contactless palmprint image faces several severe challenges, including unconstrained hand placement, complex background, light interference, etc. This paper proposes logical conjunction of triple-perpendicular-directional translation residual (TPDTR) for the improvement of contactless palmprint image preprocessing. The search of hand valley point is within the borders of hand valley gap detected by TPDTR, therefore, the computational cost is effectively decreased. Furthermore, the anti-interference capacity of region is stronger than that of point and line, so TPDTR improves the accuracy of hand valley point detection. The experimental results confirm the superiorities of TPDTR over the existing methods in computational cost and accuracy.