A novel method for real time face spoof recognition for single and multiple user authentication
S Chinchu, Anisha Mohammed, Boya Leela Mahesh · 2017
Face recognition has been a fast developing and emerging zone in the area of research over the last two decades, which involves recognition of personal identity based on statistical or geometrical features. The goal of this paper is to evaluate single and multiple user authentications in real time scenarios. Two methods have been adopted for face verification or identification. The first method uses feature descriptors such as SURF (Speeded-up robust features), Harris corner and minimum Eigen value algorithm to match faces up to a certain accuracy level and the second method is to recognize face by using a binary classifier, SVM (Support Vector Machine). The feature extraction parameters used for face recognition includes GLCM (Gray level Co-occurrences matrix), shape parameters and color features. This method was evaluated by using SVM and naïve Bayes classifier and face recognition using SVM classifier have obtained good results and among feature descriptors, Minimum Eigen corner detector outperformed better result. Finally analysis of real face and the printed photo of the same is evaluated in order to avoid face spoof attack. This paper concentrate mainly on face spoof recognition.