ROBUST CLASSIFICATION OF FACE IMAGE BASED ON SURFACE INFORMATION FOR TWO DATE-SETS ANALYSIS

The evaluation of the proposed algorithm was performed on standard data, the FRAV3D FERET datasets. Feature-selection was used as a preliminary step for a learning algorithm. The learning algorithm we used, was a simple nearest-neighbor on the selected features. We experimented with the following features: pixel values, the Viola and Jones integral images, Fourier, DCT, and Walsh Transform coefficients. The FRAV3D dataset consists of images of 20 objects. Each object has 72 images taken from different poses and there is a 5 degree pose difference between consecutive images of each object. The size of each image is 128 × 128. We evaluated the experiment using the classic 10-fold cross-validation technique. The Experimental results guarantee robustness of our proposed method.