Abstract

The process of selecting the best solution for a given problem from a large number of alternatives is known as optimization. For the face recognition system, an Improved Chicken Swarm Optimization (ICSO) algorithm was presented. Chicken Swarm Optimization is a swarm intelligence-based approach that maintains a fair balance between exploration and exploitation. Nevertheless, Standard CSO still suffers from the ease of slipping into local optimum and sluggish convergence speed when solving high dimensional issues. To better combine the global and local search, the rooster and hen position update procedure now includes a Chaotic gauss map and tent map. This is done to avoid the rooster and hens slipping into local optimum which could lead to premature convergence. For feature extraction, Local Binary Pattern (LBP) was employed, and for feature selection, the Improved Chicken Swarm Optimization (ICSO) was applied. When LBP and CSO were combined, the accuracy of the facial images was measured. When LBP and ICSO were combined, the accuracy of the facial images was measured. According to the results of the trials, LBP-CSO had a classification accuracy of 91.67%, whereas LBP-ICSO had a classification accuracy of 96.25%.