Particle Swarm Optimization vs Seed-Based Region Growing: Brain Abnormalities Segmentation

Abstract

Segmentation remains one of the most challenging problems in medical imaging due to the low contrast of images. Segmentation of Magnetic Resonance Imaging (MRI) images is an important part of brain imaging research. Segmentation can facilitate the diagnosis of neurological diseases. This paper compares the performances of Particle Swarm Optimization (PSO) and Seed-Based Region Growing (SBRG) approaches in the segmentation of human brain tissue abnormalities. In this paper, the abnormalities are categorized as “light” and “dark” according to their appearances. Fifty controlled experimental data which is also called mosaic images is used. The data is designed in such a way that prior knowledge of the size of the abnormalities are known. This is done by cutting various shapes and sizes of various abnormalities and pasting them onto normal brain tissues, where the tissues and the background are divided into “low”, “medium” and “high” intensities. The knowledge of the size of the abnormalities by the number of pixels are then used as the ground truth to be compared with both PSO and SBRG segmentation results. It was found that the proposed PSO and SBRG techniques may provide potential solutions to the current difficulties in detecting abnormalities in human brain tissue area.