A Squeeze and Excitation Framework Utilizing ResNet-152 for Alzheimer’s Disease Dementia Classification
2025 (English)In: Computational Intelligence in Pattern Recognition: Proceedings of CIPR 2024, Volume 1 / [ed] Asit Kumar Das; Janmenjoy Nayak; Bighnaraj Naik; M. Himabindu; S. Vimal, Danilo Pelusi, Cham: Springer, 2025, p. 203-213Conference paper, Published paper (Refereed)
Abstract [en]
Alzheimer’s disease, a neurodegenerative condition severely impacting cognition and memory, stands as a predominant contributor to dementia. This chapter addresses the urgent need for effective identification and categorization of Alzheimer’s Disease (AD) dementia, crucial for timely intervention and improved patient outcomes. This chapter introduces a novel Squeeze and Excitation ResNet-152 (or SE-ResNet-152) model based on Convolutional Neural Network (CNN), for the classification of Alzheimer’s disease dementia into four distinct categories namely Very Mild Dementia (VMD), Mild Dementia (MD), Moderate Dementia (MoD), and Non-demented (ND), utilizing Magnetic Resonance Imaging (MRI) images. The incorporation of the Squeeze and Excitation (SE) block plays a pivotal role by recalibrating channel-wise feature responses, thereby boosting the model’s capacity to capture informative features from input data. The SE-ResNet-152 model exhibits an impressive overall accuracy of 99.00%, emphasizing its potential as a powerful tool for accurate Alzheimer’s disease dementia classification and the subsequent facilitation of timely interventions. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.
Place, publisher, year, edition, pages
Cham: Springer, 2025. p. 203-213
Series
Lecture Notes in Networks and Systems, ISSN 2367-3370, E-ISSN 2367-3389 ; 1152
Keywords [en]
Alzheimer’s disease, a neurodegenerative condition severely impacting cognition and memory, stands as a predominant contributor to dementia. This chapter addresses the urgent need for effective identification and categorization of Alzheimer’s Disease (AD) dementia, crucial for timely intervention and improved patient outcomes. This chapter introduces a novel Squeeze and Excitation ResNet-152 (or SE-ResNet-152) model based on Convolutional Neural Network (CNN), for the classification of Alzheimer’s disease dementia into four distinct categories namely Very Mild Dementia (VMD), Mild Dementia (MD), Moderate Dementia (MoD), and Non-demented (ND), utilizing Magnetic Resonance Imaging (MRI) images. The incorporation of the Squeeze and Excitation (SE) block plays a pivotal role by recalibrating channel-wise feature responses, thereby boosting the model’s capacity to capture informative features from input data. The SE-ResNet-152 model exhibits an impressive overall accuracy of 99.00%, emphasizing its potential as a powerful tool for accurate Alzheimer’s disease dementia classification and the subsequent facilitation of timely interventions. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.
National Category
Neurology Neurosciences
Identifiers
URN: urn:nbn:se:hh:diva-55759DOI: 10.1007/978-981-97-8090-7_15Scopus ID: 2-s2.0-105000623338ISBN: 978-981-97-8089-1 (print)ISBN: 978-981-97-8090-7 (electronic)OAI: oai:DiVA.org:hh-55759DiVA, id: diva2:1951478
Conference
6th International Conference on Computational Intelligence in Pattern Recognition, CIPR 2024, Odisha, Indien, 15-16 March, 2024
2025-04-112025-04-112025-10-01Bibliographically approved