Eye-Blink Detection for Real Time Face Liveness Detection

Chao-Lung Chou ； Chung-Hsien Tsai

《TANET2016 臺灣網際網路研討會》 2016 (2016/11) Pp. 1038-1041-187

https://doi.org/10.6679/TANET.2016.187

摘要

A spoofing attack occurs as a person tries to masquerade as a valid user by presenting the fake biometric trait to the sensor. Liveness detection, or anti-spoofing, uses particular measures to reduce the vulnerability of biometric systems to the threats of spoofing attack. This paper presents an approach for real-time eye-blink detection for face liveness detection. First, the face and eye region are detection by using Viola and Jones cascaded AdaBoost algorithm. Then, the extracted image is binarized using simple addition operations to reduce the computational cost. Finally, a discrete measure is used to detect eye-blinks. Experiments indicate correct detection rate is about 88%, which indicates of the method’s simplicity and effectiveness.

關鍵字

Biometric ； Spoofing ； Liveness Detection ； Eye-blink