Gabor feature based classification using the enhanced fisher linear discriminant model for face recognition

Abstract

This paper introduces a novel Gabor-Fisher (1936) classifier (GFC) for face recognition. The GFC method, which is robust to changes in illumination and facial expression, applies the enhanced Fisher linear discriminant model (EFM) to an augmented Gabor feature vector derived from the Gabor wavelet representation of face images. The novelty of this paper comes from 1) the derivation of an augmented Gabor feature vector, whose dimensionality is further reduced using the EFM by considering both data compression and recognition (generalization) performance; 2) the development of a Gabor-Fisher classifier for multi-class problems; and 3) extensive performance evaluation studies. In particular, we performed comparative studies of different similarity measures applied to various classifiers. We also performed comparative experimental studies of various face recognition schemes, including our novel GFC method, the Gabor wavelet method, the eigenfaces method, the Fisherfaces method, the EFM method, the combination of Gabor and the eigenfaces method, and the combination of Gabor and the Fisherfaces method. The feasibility of the new GFC method has been successfully tested on face recognition using 600 FERET frontal face images corresponding to 200 subjects, which were acquired under variable illumination and facial expressions. The novel GFC method achieves 100% accuracy on face recognition using only 62 features.

Subspace Learning from Image Gradient Orientations

“…As an example, for IGO-LDA we used IGO-PCA to preserve n − C dimensions which results in the IGO-version of Fisherfaces [21]. Finally, given a test vector z, we extracted features using c = U H z and performed classification using the nearest-neighbor rule based on normalized correlation [46], [47], [14]. More details on the implementation of IGO-Methods can be found in Section IV of the accompanying supplementary material.…”

“…For Gabor features, we used the popular approach of [5], [46], [47]. In particular, we used a filter bank of 5 scales and 8 orientations and then down-sampled the obtained features by a factor of 4 (so that the number of features is reasonably large).…”

Classification of Hoi-An and Sin-Chew Agarwood by Components Analysis of VOCs Released in Heat-Treated Agarwood using TD-GCMS and Chemometric Methods

“…FDA is used to find a linear combination of continuous independent variables that characterize or separate two or more classes of objects or events (Fisher 1936;Lu et al 2017). It is a classic and popular supervised learning method commonly used in face recognition, data rating recognition, and animal and plant recognition through some main features (Liu and Wechsler 2002;Alexandre-Cortizo et al 2005;Witten and Tibshirani 2011).The Fisher criterion is defined as the ratio of the between-class variance to the within-class variance (Mahmoudi and Duman2015). Taking the standard deviation of data in both classes leads to a proper weight vector estimate to prevent overlap between the projected data in each class.…”

Fusion of Color and Modified Weber’s Law Descriptor Based Human Skin Detection

“…In recent years, more and more attention is paid to local texture features in face detection due to its more robustness to variations of facial pose, expressions, occlusion. In particular, Gabor wavelets (Liu and Wechsler, 2002) and Local Binary Patterns (LBP) which is first presented by Ojala et al (2002) are mainly used in texture detection. LBP was successfully introduced into face recognition by Ahonen et al (2006) and has gained much attention by researchers in face recognition.…”