Authors : Ismaila Folasade. M.; Afolabi O. Adeolu; Ismaila W. Oladimeji; Alo Oluwaseun O.

Volume/Issue : Volume 9 - 2024, Issue 1 - January

Google Scholar : http://tinyurl.com/umm9z7xk

Scribd : http://tinyurl.com/y8e7w7t4

DOI : https://doi.org/10.5281/zenodo.10670429

Abstract : Digital image processing is the use of computer algorithms to analyze digital images. Digital image processing, involves many processing stages of which feature extraction stage is important. Feature extraction involves reducing the number of resources required to describe a large set of data. However, choosing a feature extraction techniques is a problem because of their deficiencies. Thus, this paper presents a comparative performance analysis of selected feature extraction techniques in human face images. 90 face images were acquired with three different poses viz: normal, angry and laughing. The face images were first pre-processed and then subjected to selected feature extraction techniques (Local binary pattern, Principal component analysis, Gabor filter and Linear discriminant analysis). The extracted features were then classified using Backpropagation neural network. The results of recognition accuracy produced by Gabor filter, PCA, LDA and LBP at 0.76 threshold are 76.7%, 72.2%. 78.9% and 85.6%. Hence, it can be deduced that LBP performed the best among the four selected feature extraction techniques.

Keywords : Digital Image Processing, Feature Extraction, Local Binary Pattern, Principal Component Analysis, Gabor Filter, Linear Discriminant Analysis.