Design of Face Recognition System Realized with the Aid of PCA-Based RBFNN
Sun-Hwan Kim, Sung‐Kwun Oh, Jin-Yul Kim · 2016
In this study, we propose a robust face recognition method that is based on pose estimation using Fuzzy C Means-based Radial Basis Function Neural Networks(RBFNNs) that consist of three functional modules. Generally face recognition is conducted when object looks at the front, and this situation to limit the application range of face recognition often occurs in reality. To overcome these restriction, we introduce the procedure for pose estimation that use a similarity calculated with trained pose model by using the eigenfaces of Principal Component Analysis(PCA) in various poses. We make an experiment for its performance by using Cambridge Head Pose database and our university's IC&CI lab face database.