Alzheimer’s Disease MRI Brain Segmentation Using Pythagorean Fuzzy Sets

Alzheimer’s disease (AD) is a degenerative and ultimately fatal brain disorder for which there is no cure. This neurological condition, with a complex etiology, causes dementia and cognitive decline, making its identification challenging due to the variation in brain MRIs, including differences in s...

Full description

Saved in:
Bibliographic Details
Published in:Traitement du signal 2024-10, Vol.41 (5), p.2701-2709
Main Authors: Latha, Vunnam Asha, Namburu, Anupama
Format: Article
Language:eng ; fre
Subjects:
Algorithms
Alzheimer's disease
Brain
Centroids
Cerebrospinal fluid
Clustering
Cognitive ability
Decision making
Dementia
Fuzzy sets
Image processing
Image segmentation
Medical imaging
Neuroimaging
Uncertainty analysis
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Alzheimer’s disease (AD) is a degenerative and ultimately fatal brain disorder for which there is no cure. This neurological condition, with a complex etiology, causes dementia and cognitive decline, making its identification challenging due to the variation in brain MRIs, including differences in size, shape, and CSF flow. While there is no treatment for AD, its progression can be slowed with early diagnosis. Many researchers have employed image processing-based techniques to differentiate between normal and AD-affected patients based on brain images. However, the brain's regions often look super similar, making it tricky to pinpoint specific areas, plus there's always some uncertainty when it comes to extracting the exact regions. There have been various proposals in the literature for fuzzy c-means and intuitionistic fuzzy c-means (IFCM) approaches to deal with this ambiguity and uncertainty. In contrast, Pythagorean fuzzy sets (PFS) provide a more precise means of verifying membership, making them an effective tool for managing uncertainty. The author analyzed PFS and applied fuzzy c-means to propose Pythagorean fuzzy c-means (PFCM). Additionally, histogram-based initial centroids were used to avoid the local minima problem, which is common in many clustering algorithms. The proposed clustering algorithm demonstrated improved performance, completing execution in less than 1.5 seconds, owing to the incorporation of initial centroids and PFS-based clustering. The proposed method achieved high accuracy rates: 98.64% for white matter (WM), 97.4% for gray matter (GM), and 98.14% for cerebrospinal fluid (CSF) in detecting brain tissues.
ISSN:0765-0019
1958-5608
DOI:10.18280/ts.410543