Global and Local Classifiers for Face Recognition

Face recognition means checking for the presence of a face from a database that contains many faces. The face images considered for recognition undergo large variations due to changes in illumination conditions, viewing direction, facial expression and aging etc. The face images have similar geometrical features and hence discriminating one face from the other in the database is a challenging task. Global and Local features are crucial for face recognition. In the proposed method, both the global and local features are extracted from the input face images. Global features are the holistically structural configuration of facial organs, as well as facial contour. Global Features are extracted from whole face images by keeping the low frequency coefficients of Fourier transform. Real and imaginary components in the low frequency band are concatenated into a single feature vector named Global Fourier Feature Vector (GFFV). Local features are high frequency and dependent on position and orientation of the face images. Local features are extracted by Gabor wavelets. Gabor features are spatially grouped into a number of feature vectors named Local Gabor Feature Vector (LGFV). Fisher’s Linear Discriminant (FLD) is separately applied to the Global Fourier Features and each local patch of Gabor features. The resultant vectors are fused using region based fusion algorithm. The processed test face image is verified for a match with the faces in the database and recognition is done.