The Research of the Liveness Detection Method Based on Sparse Support Vector Machines

Lijun Yu,Ying Ding,Hui Wang,Weijie Sun
DOI: https://doi.org/10.1109/icma.2018.8484438
2018-01-01
Abstract:With the development of face detection technology in identity authentication system, human face detection technology is becoming more and more important. It is very easy to use photographs and videos and other ways to get fake face to attack the identity authentication system. This paper designs a method on judging the authenticity of faces based on the texture feature difference between real face images and deception photos. The system is constructed by VC++ and OpenCV, mainly composed of image reading, face and facial features detection, image block feature extraction and classification of the support vector machine. In this paper, sparse support vector machine is used to train and classify data, and the application of sparse support vector machine in liveness detection is the innovation point of this paper. The results of experiments on a specific database show that it has good performance in terms of accuracy and speed.
What problem does this paper attempt to address?