PCA is a common feature extraction and data representation method which is used mainly for dimensionality reduction in pattern recognition area. However, this method is usually affected by light illumination. Meanwhile, LDA is another popular dimensionality reduction technique that achieves maximum discrimination by finding the maximum ratio of the between to the within class distance. In this work, the performance of face and ear recognition using SVM based on pca and SVM based on a combination of PCA and LDA techniques have been examined with the PCA and LDA techniques depend on distance similarity measures. The experimental work shows that the recognition performance based on SVM_pca and SVM_pca,lda techniques outperform the recognition performance based on PCA using distance similarity measures. While, No significant differences were found using SVM_pca, SVM_pca,lda and LDA based on distance similarity measures. The experimental results using accuracy and the error rate indicated that the recognition depending on SVM_pca and SVM_pca,lda techniques lead to a best performance than using PCA based on distance similarity measures….
