Automatic Seeded Region Growing (ASRG) Using Genetic Algorithm for Brain MRI Segmentation

Segmentation images into several parts are an important task in many image processing and computer image applications. The purpose of dividing a photo into several parts involves segmentation the image into different regions based on the criteria for future processing. One of the methods of segmentation the images is the growth method of the area. In this study, the region’s growth method is used to segment the brain MRI images. The method of growing the area consists of several steps. In the beginning, you have to select a few initial points (seeds) that are related to the areas to be separated from the field. Then, starting from these points in the neighborhood of these points, checked the other points and if added according to the selectivity similarity criterion belonging to the first point region. In this research, the selection of initial points is done automatically. For this purpose, the genetic algorithm is used which searches for the optimal initial points using the initial population selection and the definition of the proper fitness function. Investigating the result of the proposed method on the images compared with the region’s growth method by manually selecting the initial points indicates that this method can well improve the segmentation of the brain MRI image.