Performance evaluation of Alzheimer’s disease detection in brain MRI images using the Alzheimer-net classifier

• Published in: Engineering Research Express Vol. 7; no. 3; pp. 35209 - 35221
• Main Authors: T, Prasath, V, Sumathi
• Format: Journal Article
• Language: English
• Published: IOP Publishing 30.09.2025
• Subjects: classifiers; deep learning; gaussian noise; MRI images
• ISSN: 2631-8695; 2631-8695
• DOI: 10.1088/2631-8695/ade8d5

Introduction : Alzheimer’s disease is a neurologisscal condition that affects the elderly people, which in turn leads to memory loss.Early detection helps to slow down the disease progression. Methods : Using a modified deep learning methodology, this study suggests an effective workflow for classifying the Alzheimer’s disease (AD) from the Non –Alzheimer’s disease (NAD) image. Both AD and NAD pictures were pre-processed in this work to remove Gaussian noise, and the noise-removed image was subsequently divided into a number of decomposed coefficient sub-bands using the Non-Sub sampled Contour let Transform (NSCT). To effectively classify the AD image from the NAD images, features from these sub-bands have been identified and incorporated into a modified deep learning system. Two datasets namely Kaggle and MIRIAD were used to evaluate the model performance. Results : The proposed approach achieves 97.9% AD Classification Index (ADCI) for the Kaggle dataset and 97.9% ADCI for the MIRIAD dataset. Also this approach achieves an 98.7% Non-AD Classification Index (NACI) for the Kaggle dataset and 97.1% NADCI for the MIRIAD dataset in the NAD detection. Conclusions : In this investigation, the outcomes have been evaluated and verified with regard to other comparable modelling techniques.