C050

Finger Vein Presentation Attack Detection using Local Directional Pattern

Nurul Nabihah Ashari, Ong Thian Song, Tee Connie

AFFILIATION
Faculty of Information Science & Technology, Multimedia University

Description of Invention

In the recent years, finger vein biometrics has been gaining traction commercially and is susceptible to presentation attacks. Presentation attacks is one where the attacker uses fake finger vein pattern to spoof the finger vein sensor. In this research, Local Directional Pattern is proposed for presentation attack detection (PAD). The best ACER result for SCUT dataset is 4.26% and 4.66% for VERA dataset.