A Novel Hybrid Model for Multi-Class AD Classification using OASIS MRI Dataset
International Journal of Experimental Research and Review · 30.04.2026 · 50 · 25-40
DOI: 10.52756/ijerr.2026.v50.002 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Experimental Research and Review, online erschienen am 30.04.2026. Verfasst von Shagufta Shakeel und 1 weiteren Autor:innen. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Alzheimer's disease (AD) is a chronic and progressive neurodegenerative disease and the commonest type of dementia? The disorder is characterized by gradual loss of cognition and memory? Early diagnosis and recognition of AD is vital for treatment? Diagnosis is very difficult since there are minimal and subtle structural changes in the brain that can be seen through MRI scans? Although many DL-based methods have been developed with good outcomes‚ they do not work well with generalization‚ single-modal data and multi-class labels? To classify the AD stages‚ it is imperative to develop generalizable DL models that can learn from neuroimaging data the features that are distinctive at different stages of disease progression? This paper develops a DL-based multi-class classification model using the OASIS MRI dataset? The model pre-processes the grayscale image‚ spatially normalizes it‚ and extracts ROI by performing thresholding and morphological processing to focus on relevant regions? Afterward‚ a compact CNN and a Transformer encoder are used to extract and differentiate the image features in different stages by learning their global context? Stratified 10-fold cross-validation is used with balanced batch sampling for both training and computing the performance metrics? The accuracy‚ precision‚ recall‚ and F1 score of 90?46%‚ 91?05%‚ 90?46%‚ 90?54% respectively have been achieved by the proposed model in the conducted experiments showing very high accuracy and a consistent performance with high precision‚ recall and F1 score? The model shows content validity and the ability to differentiate between stages of AD effectively? Future research may continue by employing multimodal data and investigating more complex hybrid models? The proposed model will be used for early diagnosis of the disease and as a part of a decision support system? Furthermore, we have also compare our proposed model with some existing models that worked on OASIS dataset and found that our model outperforms the existing model.
Zitieren
APA
Shakeel, S., & Yadav, S. P. (2026). A Novel Hybrid Model for Multi-Class AD Classification using OASIS MRI Dataset. International Journal of Experimental Research and Review, 50, 25-40. https://doi.org/10.52756/ijerr.2026.v50.002
Vancouver
Shakeel S, Yadav SP. A Novel Hybrid Model for Multi-Class AD Classification using OASIS MRI Dataset. International Journal of Experimental Research and Review. 2026;50:25-40. doi:10.52756/ijerr.2026.v50.002
BibTeX
@article{Shakeel2026,
author = {Shakeel, Shagufta and Yadav, Satya Prakash},
title = {A Novel Hybrid Model for Multi-Class AD Classification using OASIS MRI Dataset},
journal = {International Journal of Experimental Research and Review},
year = {2026},
volume = {50},
pages = {25--40},
doi = {10.52756/ijerr.2026.v50.002}
}
EUVIMED