Acute Lymphoblastic Leukemia Classification

This research introduces an intelligent decision support system for diagnosing acute lymphoblastic leukemia (ALL) through microscopic images, employing two modified Bare-bones Particle Swarm Optimization (BBPSO) algorithms to effectively classify ALL by identifying key characteristics of healthy and blast cells. The first BBPSO variant enhances the search process with accelerated chaotic behaviors such as food chasing and enemy avoidance, preventing premature convergence. Both algorithms were tested against the ALL-IDB2 database, demonstrating outstanding classification performances with geometric mean accuracies of 94.94% and 96.25%, significantly surpassing other metaheuristic search methods in ALL diagnosis.