Related Experiment Video
Updated: Feb 22, 2026

Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
Published on: January 7, 2019
A Metaheuristically Tuned Interval Type 2 Fuzzy System to Reduce Segmentation Uncertainty in Brain MRI Images
Abolfazl Taghribi1, Saeed Sharifian2
1Department of Electrical Engineering, Amirkabir University of Technology, Tehran, 15914, Iran.
Abstract:
Precise segmentation of magnetic resonance image (MRI) seems challenging because of the complex structure of the brain, non-uniform field in images, and noise. As a result, decision-making is associated with uncertainty. Fuzzy based approaches have been developed to overcome this problem, though most of them use fuzzy type 1 method, and sometimes contain a pre-processing step. This paper "modified type 2 fuzzy system" (MT2FS) declares a state-of-the-art method to segment MRI images using interval fuzzy type-2. Furthermore, Genetic algorithm has been employed to specify the best values for mean and variance of upper and lower membership functions. This strategy will determine discrimination boundaries for different brain tissues to be less independent from the training set. Finally, the result of fuzzy rules is extracted by using Dempster-Shafer rule combination method. Simulation results demonstrate a satisfactory output on both simulated and real MRI images in comparison with previously conducted research works without the need for a pre-processing stage.

