Face recognition using Gabor Wavelet and Non-negative Matrix Factorization
Fredy Purnomo, Derwin Suhartono, Muhsin Shodiq, Albert Susanto, Steven Raharja, Ricky Wijaya Kurniawan · 2015
Biometrics authentication is commonly used as a tool to increase security level. Face recognition is an example of authentication by using biometrics. Face recognition requires specific methods to obtain face representation as its features. There are many methods which have been developed to get these unique features, such as PCA, LDA, ICA and hybrid methods like ICA and SVM, Gabor and ICA and many others. This research develops a hybrid method from Gabor Wavelet and Non-negative Matrix Factorization (NMF) because during conducting literature reviews, the combination of these methods has never been found. The objective of this research is to create face recognition method with a better accuracy than the previous methods. Testing is conducted by using ORL face database and it gets around 95% accuracy rate. This result shows that the proposed method indicates a better accuracy compared with the previous methods.