Identifying Dementia in MRI Scans Using Artificial Neural Network and K-Nearest Neighbor

  • Abhinash Manandhar Department of Electronics and Computer Engineering, IOE Pulchowk
  • Sagar Gautam Department of Electronics and Computer Engineering, IOE Pulchowk
  • Dharma K. Shrestha Department of Electronics and Computer Engineering, IOE Pulchowk
  • Sujan Sauden Department of Electronics and Computer Engineering, IOE Pulchowk
  • Dibakar R. Pant Department of Electronics and Computer Engineering, IOE Pulchowk

Abstract

Artificial Neural Network (ANN) and K-Nearest Neighbor (KNN) models are used to detect dementia in MRI scans. Each scan is segmented and then normalized into four discrete color areas corresponding to white matter, dark gray matter, light gray matter and black background, the former three constituting the feature set. Thresholding technique is used in segmentation and nearest neighbor interpolation technique is used in normalization of the images. The ANN implementation resulted into 69.81% accuracy and the KNN implementation resulted into 81.13% accuracy in classification of demented and non-demented scans. 

Published
Dec 21, 2016
How to Cite
MANANDHAR, Abhinash et al. Identifying Dementia in MRI Scans Using Artificial Neural Network and K-Nearest Neighbor. Zerone Scholar, [S.l.], v. 1, n. 1, p. 22-25, dec. 2016. ISSN 2542-2774. Available at: <http://zerone.pcampus.edu.np/ojs/index.php/scholar/article/view/7>. Date accessed: 28 may 2022.