Face recognition system using adaptive neurofuzzy inference system

Biometric verification using Face Recognition (FR) is one of the most emerging trait in the field of image analysis. The steps involved in FR are classified into two types: 1) face verification and 2) face classification. In recent years, the face verification systems are well-developed, but still face classification algorithms are facing problems like poor outcome in severe lighting variations, illuminance, face pose variation, etc. In order to overcome these difficulties, we propose Adaptive Neuro-Fuzzy Inference System (ANFIS) with Principal Component Analysis (PCA) by considering different contributions of the training samples (ORL and YALE B dataset). At first, the facial images are enhanced by using DualTree Complex Wavelet Transform (DTCWT) approach. The preprocessed facial image features are extracted by employing PCA method. Using the feature information, the facial classification is achieved by using ANFIS classifier. Experimental outcome shows that the proposed approach improved accuracy in face detection up to 0.2-0.8% compared to the neural network (existing method).