The impact of image block size on face feature extraction using Discrete Cosine Transform

Abstract

In this paper we conduct an experiment to study the effects of multiple block sizes in face images using the Discrete Cosine Transform (DCT) algorithm. Facial features are extracted from each block using the DCT algorithm. These features are then combined to form a feature vector for facial recognition. The goal of the paper is to discover if there is an underlying principle for determining the best block size for increasing the recognition accuracy with the DCT, when it is being used for facial recognition. The support vector machine (SVM) algorithm is used for facial recognition experiments.