MINIMUM CROSS ENTROPY BASED IMAGE SEGMENTATION USING NEW HEURISTIC OPTIMIZATION TECHNIQUE

Image thresholding is an important technique for image processing and pattern recognition. Multilevel thresholding problem is often treated as a problem of optimization of an objective function. In this paper, minimum cross entropy (MCE) is introduced for multilevel thresholding which uses Improved Bacterial Foraging (IBF) algorithm for minimizing the MCE objective function. Some examples of test images are presented to compare the segmentation methods based on the IBF approach, with bacterial foraging (BF) algorithm, particle swarm optimization (PSO) algorithm and genetic algorithm (GA). From the viewpoint of visualization and image contrast, experimental results show that the thresholding method based on IBF method performs better than the BF, PSO and GA method. Moreover, the proposed method provides better accuracy, stability and computational efficiency.