원문정보
초록
영어
In recent years, counterfeiting is posing a serious threat to face recognition based authentication system. Introducing liveness detection into a face-based authentication system can be a good solution to the problem. This paper applied eye region detection and tracking technology to anti-counterfeiting application, and designed an eye region activity based liveness detection system. First, a scale balance and a “face-to-eye” strategy were proposed to fast detect the eye region, and then KLT algorithm was used to track the detected eye region. At last, a random instruction (closing eye, moving to the left or right) was evoked to the user, and the liveness judgment was given according to the results of the users’ cooperation. Experiments show that the proposed eye region activity based face liveness detection system is real-time and robust, and can meet the needs of practical applications.
목차
1. Introduction
2. Eye Region Search with Modified Viola-Jones Detection Algorithm
2.1 Face Detection with Cascade-Structured AdaBoost Algorithm
2.2 Using Scale Balance Strategy to Accelerate Face Detection
2.3 ‘Face-Eye’ Searching Strategy for Eye Region Detection
3. KLT Tracking Algorithm
3.1 KLT Tracking Algorithm
3.2 Eye Region Tracking based on KLT Algorithm
4. Eye Region Activity State Based Face Liveness Detection
4.1 State of Eyes closed Detection Algorithm based on NCC
4.2 The Implementation of Face Liveness Detection System
5. Summary
Reference