Application of ant colony optimization for image segmentation

Ant colony optimization application for two-dimensional electrophoresis gel image segmentation is investigated. Standard ACO model with one ants' colony is presented also provided ants' colony size control method. Presented modified ACO model for two-dimensional electrophoresis gel image segmentation. Created synthetic image and experimentally estimated model parameters for segmentation are provided. Image pre-processing operations and processing results for decrease of processing time and increase segmentation accuracy are presented. Obtained results show that model is suitable for overlapped protein spots segmentation. Was reached average segmentation result that depends on number of overlapped spots and how close they centers are to each other. Model parameters and their influence on ant's behaviour and segmentation results are provided. Segmentation dynamics analyzed and intermediate segmentation results are provided. Population changes analyzed and population size change is proposed as a stopping parameter for segmentation process. Segmentation accuracy was about 67% improve over simple threshold function segmentation. Processing times are commented. Segmentation result can be improved by the use of real electrophoresis images for parameters estimation, also by the use of more effective optimal model parameters estimation algorithm.