Automated Brain Image Segmentation

Syahirah Husna, Uda Zahli (2014) Automated Brain Image Segmentation. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Automated Brain Image Segmentation 24 Pages.pdf - Submitted Version Download (780kB)

Abstract

This project is about image segmentation application to automatically segment the Magnetic Resonance Imaging (MRI) brain image using thresholding and fuzzy c-means (FCM) methods. Conventional method of brain MRI segmentation is done manually by neuro-radiologists which are time consuming and have significant differences between expertises. Therefore, automatic segmentation would serve as second option with neuro-radiologists. The objective of the project is to analyze the automated brain MRI image segmentation by using thresholding and FCM clustering. Thresholding technique identifies a region based on the pixels with similar intensity values and provides boundaries in images that contain solid objects on a contrast background. This technique gives binary output image from a grey scale image. FCM is an iterative process. The iteration is being repeated until a set point called the threshold is reached or the process stops when the maximum number of iterations is reached. The steps include in the project are pre-processing, segmentation (thresholding and FCM) and performance analysis. FCM method is more accurate compare to the thresholding which the percentage of accuracy is 45.45% compare to 37.50 %.

Actions (login required)

View Item