EDGE DETECTION IN MRI BRAIN TUMOR IMAGES USING MODIFIED ACO ALGORITHM BASED ON WEIGHTED HEURISTICS

Images play a crucial role in the medical field across various aspects, from diagnosis to treatment planning. The Magnetic Resonance Imaging (MRI) brain tumor images are always associated with noise. Therefore, to diagnose diseases, image edge detection plays a challenging role in the medical field. It identifies edges, boundaries, disruptions, irregularities, and other valuable features. Ant colony optimization (ACO) is a well-known metaheuristic algorithm inspired by the way ants lay down a chemical pheromone while searching for food. It generates a pheromone matrix, which provides edge information accessible at every pixel of the image, developed by ants navigating across the image. The movements of ants depend on the local variance of the image intensity value. Conventional statistical range-based approaches are limited in accurately identifying weak edges. The proposed approach enhances edge detection using ACO, integrating weighted statistical range-based heuristics information and Gaussian gradients to generate a binary image that enables to detect the strong edges. Thus by assigning weights to the neighborhood range of pixels helps to determine the direction in which ants are able to move. The Gauss gradient produces the edges effectively. The proposed method was tested with standard MRI brain tumor images. Experimental results exhibit the comparable performance of the proposed method with conventional edge detectors in terms of performance parameters like Figure of Merit, Sensitivity, Accuracy, and output images.