1) The document proposes four techniques for facial feature extraction - 2DPCA, LDA, KPCA, and KFA. It uses these techniques to extract features from images in the ORL database, applies LDA for classification, and recognizes faces using a feedforward neural network (FFNN). 2) It compares the recognition rates of the different feature extraction techniques when using 10, 20, 30, and 40 hidden neurons in the FFNN. The technique that performs best is KPCA, achieving a maximum recognition rate of 94.82% using 20 hidden neurons. 3) In conclusion, the goal of implementing and comparing different feature extraction techniques with an FFNN classifier on face images was achieved. KPCA provided the best