Hybrid heuristics for mammogram segmentation

Simulated Annealing (SA), Tabu Search (TS), Genetic Algorithm (GA), and Ant Colony System (ACS) are four of the main algorithms for solving challenging problems of intelligent systems. In this paper, these four techniques and three novel hybrid combinations of them are proposed to mammogram segmentation. The novel hybrid algorithms consist of a Sequential TS-ACS, a Hybrid ACS/TS, and a Sequential ACS-TS algorithms. Initially the mammogram images are enhanced and Markov Random Field (MRF) is applied to label the image pixels, and the Maximizing a Posterior (MAP) is calculated for each pixel. To find out the optimum label, which minimizes the MAP estimate, the heuristic algorithms are applied. Statistical comparative analysis conclude that all of the three proposed novel techniques are significantly better than each of their nonhybrid competitors, and furthermore the Sequential ACS-TS provides the superior solution of all.