An effective approach for white matter, grey matter, and cerebrospinal fluid segmentation from 3D brain MRI. (8th September 2021)

• Record Type: Journal Article
• Authors: Bao, Pham The; Tran, Tuan; Tuan, Tran Anh; Thuy, Le Nhi Lam; Kim, Jin Young
• Is Part Of: International journal of intelligent information and database systems. Volume 14:Number 4(2021)
• Page Start: 315
• Page End: 332
• Publication Date: 2021-09-08
• Subjects: medical imaging -- image segmentation -- brain MRI segmentation -- tiny 3D U-Net -- 2D U-Net; Database management -- Computer programs -- Periodicals; Information retrieval -- Computer programs -- Periodicals; Information storage and retrieval systems -- Computer programs -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Intelligent agents (Computer software) -- Periodicals; 006.33
• Languages: English
• ISSNs: 1751-5858

Abstract

According to the World Alzheimer Report 2019, over 50 million people are living with dementia globally and it will increase to 152 million by 2050. Alzheimer's disease (AD) is a common type of dementia. One of the methods for Alzheimer's disease detection is using medical imaging especially using magnetic resonance imaging (MRI) in the brain concerning the white matter (WM), grey matter (GM), and cerebrospinal fluid regions (CSF). There are many methods based on a 3D CNN model for tissue segmentation but the disadvantage of the 3D model compared with the 2D model is the number of training objects will be decreased and the computation cost is increased. In this research, we proposed an effective approach to segment automatically WM, GM, and CSF in 3D brain MRI by using 2D and 3D CNN model combination. Our proposed method was evaluated with IBSR 18 and compared against state-of-the-art methods. The accuracy is 0.93, 0.92 and 0.79 for WM, GM, and CSF segmentation.