• Ravindra Dabhade

Eigenspace based face recognition systems have been most widely used flourishing approach in computer vision in recognition of faces from digital images. Numerous different eigenspace based approaches has been proposed up to now starting with the Eigenface algorithm due to its simplicity less and computational cost. In any face recognition system usually the face images to be identified are different from those stored in the database which causes the differences in the feature set too. However any victorious face recognition method depends mainly upon the selection of important features utilized by classifier. Feature selection within face recognition engross the derivation of most important features from raw input data so as to decrease the quantity of data used for classification and concurrently provide enhanced discriminatory capability. The most widely used and popular approaches for choosing the subset of features is Principle component analysis (PCA) and linear discriminant analysis (LDA). However, due to the difficulties such as lighting, expression and pose variations, these systems, individually or in combination does not present acceptable accuracies under additional realistic operating conditions. So we propose here a new method of feature extraction which consists of Sampled Gabor Features followed by PCA for dimension decrease and LDA for feature discrimination. Finally for classification purpose we suggest Cosine distance classifier.

Face recognition,Gabor filter ,distance metric ,feature extraction

"Face recognition with Gabor features, PCA, LDA, and cosine distance classifier", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 5, page no.498-506, May-2019, Available :http://www.jetir.org/papers/JETIR1905P75.pdf

"Face recognition with Gabor features, PCA, LDA, and cosine distance classifier", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 5, page no. pp498-506, May-2019, Available at : http://www.jetir.org/papers/JETIR1905P75.pdf