Improved Feature Selection in Bacteria Detection Using Two-phase Mutation Grey Wolf Optimizer

Traditional bacterial identification methods relying on microscopic and biological analyses are often constrained by high time, cost, and labor requirements. To address these limitations, this study proposes an enhanced bacterial detection framework integrating the Two-phase Mutation Grey Wolf Optimizer (TMGWO) for feature selection. The TMGWO algorithm is employed to identify the most informative feature subsets, minimize redundancy, and improve classification performance. The proposed workflow comprises data collection, preprocessing, segmentation, feature extraction, feature selection, and classification using the K-nearest neighbor (KNN) algorithm. Experimental evaluation demonstrates that incorporating TMGWO improves classification accuracy by 6.14% and reduces execution time by 33.6 seconds compared with models that exclude feature selection. These results confirm that the TMGWO-based approach offers a robust and efficient solution for automated bacterial detection in microscopic imagery.