Face Recognition Using Generalized Pseudo-Zernike Moment

Feature extraction is one of the important tasks in face recognition. Structural and statistical based approaches are two broad categories of feature extraction. This paper proposes a statistical approach for feature extraction based on Generalized Pseudo-Zernike Moment (GPZM) invariants which is powerful to characterize the image using region based shape features and also invariant to size, tilt, rotation and insensitive to noise. To achieve face recognition with higher performance the extracted features are recognized using Radial Basis Function Neural Network Classifier. The proposed method is tested using YALE database. Experimental results show that the proposed method outperforms Zernike and Pseudo-Zernike moments both in noise free and noisy conditions.