Comparative Analysis of Selected Evolutionary Algorithms as Feature Selectors in Digital Face Image Processing

Abstract

Feature selection aims to choose a small subset of the relevant features from the original ones by removing irrelevant, redundant, or noisy features. Nowadays, researchers employ evolutionary algorithms (these are efficient heuristic search methods based on Darwinian evolution with powerful characteristics of robustness and flexibility to capture global solutions of complex optimization problems) for feature selection in images classifications. However, the performances of these evolutionary algorithms varies in image processing, hence the best algorithm cannot be ascertain. Thus, this paper carried out a performance evaluation on some selected feature selector based evolutionary algorithms (Ant colony optimization, Gravitational Search algorithm, Particle Swarm Optimization and Firefly algorithm). 120 face images were collected and pre-processed to remove irrelevant features. The pre-processed images were subjected to each of the selected feature selectors to select salient features. Image matching was done with Back-propagation neural network. The results showed that Gravitational Search algorithm outperformed other techniques with accuracy of 88.3% while Particle Swarm Optimization gave 75.8%, Ant colony optimization produced 76.7% and Firefly algorithm generated 80.8%.