Alzheimer Disease Detection and Classification Using NASSNet Mobile Network

Abstract

This study investigates Alzheimer's disease (AD) detection using the NASNet Mobile architecture, emphasizing transfer learning and dataset augmentation. The baseline model achieved 94?curacy, while our enhanced approach demonstrated superior performance, reaching 100?curacy. A key improvement involved expanding the dataset from 6,400 to 12,800 images by integrating data from diverse sources, including the Alzheimer's Disease Neuroimaging Initiative (ADNI) and Open Access Series of Imaging Studies (OASIS). The augmented dataset addressed prior limitations in class representation, particularly for moderate dementia cases. Our methodology incorporated advanced preprocessing techniques and a systematic train-test-validation split to optimize model training and evaluation. The experiments confirmed the efficacy of transfer learning in leveraging pre-trained models for medical image classification tasks. Comparative analysis highlighted the significant impact of dataset size and diversity on model accuracy, with our approach achieving the highest reported performance. This research underscores the potential of deep learning architectures like NasNetMobile in improving diagnostic precision for AD across all stages, offering valuable insights for early detection and intervention. The results validate the model's robustness and demonstrate its promise for deployment in clinical settings to enhance patient care and outcomes in neurodegenerative disease management.

Keywords

Introduction

Alzheimer disease is a progressive neurodegenerative disorder, constituting the majority of dementia cases worldwide. Initially identified by Dr. Alois Alzheimer in 1906, the disease is characterized by the accumulation of beta-amyloid plaques and tau tangles in the brain, leading to the gradual deterioration and death of neurons. The stages of Alzheimer's range from preclinical, where symptoms are not apparent, to severe dementia, marked by a profound impact on both mental and physical abilities. The stages include mild cognitive impairment, mild dementia, moderate dementia, and severe dementia. Each stage exhibits specific symptoms, such as memory loss, impaired judgment, personality changes, and eventually, a decline in communication and physical capabilities (Aging, 2003). Alzheimer progression varies among individuals, making the stages rough generalizations. The disease often begins silently in the preclinical stage, with symptoms emerging gradually. Mild cognitive impairment involves subtle memory and thinking changes, while mild dementia is marked by significant memory and judgment issues. As the disease advances to moderate dementia, confusion deepens, and assistance with daily activities becomes necessary. In the severe dementia stage, mental function further declines, impacting movement and physical abilities, ultimately leading to the loss of communication and the need for full-time care. The continuous and diverse nature of Alzheimer's underscores the importance of personalized understanding and care for individuals affected by this challenging condition (States & Alzheimer, 1906). Alzheimer disease is a complex condition that affects millions of people worldwide, and its impact extends beyond the individuals diagnosed to their families and caregivers. As our understanding of Alzheimer's has deepened over the years, researchers have identified various risk factors and potential mechanisms underlying the disease's progression (Ghazal et al., 2022). One significant aspect of Alzheimer pathology is the accumulation of abnormal protein deposits in the brain. Beta-amyloid plaques and tau protein tangles are hallmark features of the disease, contributing to neuronal damage and cognitive decline. While these protein aggregates are central to Alzheimer pathology, researchers continue to investigate the precise role they play in the onset and progression of the disease (Chaihtra & Vijaya Shetty, 2021). In addition to protein abnormalities, inflammation and oxidative stress are believed to contribute to neuronal damage in Alzheimer. The brain's immune response may become dysregulated, leading to chronic inflammation that exacerbates neurodegeneration. Oxidative stress, resulting from an imbalance between free radicals and antioxidant defenses, can also contribute to cellular damage in the brain (Ashraf et al., 2021). Genetic factors also play a significant role in Alzheimer's disease. While most cases are sporadic, meaning they occur without a clear familial pattern, a small percentage of cases are linked to specific genetic mutations. For example, mutations in genes such as APP (amyloid precursor protein), PSEN1 (presenilin 1), and PSEN2 (presenilin 2) are associated with early-onset familial Alzheimer's disease (Wright, 2010). However, genetic predisposition is just one piece of the puzzle. Environmental and lifestyle factors, such as cardiovascular health, education level, and social engagement, also influence Alzheimer's risk. For instance, conditions like hypertension, diabetes, and obesity have been linked to an increased risk of developing Alzheimer's later in life (Johnson et al., 2012). Conversely, maintaining a healthy lifestyle, including regular physical activity, cognitive stimulation, and a balanced diet, may help reduce the risk or delay the onset of cognitive decline. The clinical manifestation of Alzheimer's disease varies widely among individuals, leading to a diverse range of symptoms and disease trajectories. While memory loss and cognitive decline are hallmark features, other symptoms such as changes in mood, behavior, and language abilities can also occur. The progression of Alzheimer's is often characterized by a gradual worsening of symptoms, eventually leading to severe cognitive impairment and functional dependency (Johnson et al., 2012). Given the complex nature of Alzheimer's disease, personalized approaches to diagnosis, treatment, and care are essential. Early detection through cognitive testing and neuroimaging techniques allows for timely intervention and support for individuals and their families (Johnson et al., 2012). While there is currently no cure for Alzheimer's, various pharmacological and non-pharmacological interventions aim to alleviate symptoms, improve quality of life, and provide support for caregivers. Research into novel therapeutic strategies, including disease-modifying drugs and targeted interventions, continues to advance our understanding of Alzheimer's disease (Johnson et al., 2012). Clinical trials investigating potential treatments are underway, offering hope for future breakthroughs in disease management and prevention (Johnson et al., 2012). Alzheimer disease (AD) poses a significant global health challenge, impacting individuals, families, and society. Effective strategies for prevention, diagnosis, and treatment rely on a deep understanding of the disease's underlying mechanisms, risk factors, and clinical presentation (Saleh et al., 2023). Current Alzheimer disease detection and classification methods face limitations stemming from dataset constraints, including insufficient size and diversity. This lack of diversity hinders the scalability and generalizability of models across different patient populations and imaging protocols, particularly concerning the limited representation of moderate dementia (Mod.D) cases. Although advancements have been made, such as Sharma et al.'s (2022) modified Inception model utilizing transfer learning (TL), further enhancements are necessary. This study addresses these limitations by proposing a comprehensive approach incorporating several key improvements. First, oversampling techniques are employed to augment existing datasets, mitigating the effects of class imbalance and improving model robustness. Second, the study integrates larger and more diverse datasets, including the Alzheimer's Disease Neuroimaging Initiative (ADNI) and Open Access Series of Imaging Studies (OASIS), to enhance the model's ability to generalize across different populations and imaging modalities. Third, advanced deep learning architectures, such as NasNetMobile, are utilized for both detection and classification tasks. These architectures are known for their efficiency and performance, making them suitable for clinical applications. A key focus of this research is the novel application of transfer learning to explore knowledge transfer between different Alzheimer disease stages. Specifically, the study investigates the transfer of knowledge from the Mild and Moderate stages to the earlier stages of Very Mild Alzheimer disease. This targeted approach aims to improve early detection capabilities, which is crucial for timely intervention and improved patient outcomes. By leveraging knowledge learned from more advanced stages of the disease, the model can potentially identify subtle changes indicative of very early Alzheimer disease, which are often difficult to detect using traditional methods. This transfer learning strategy, combined with dataset augmentation and advanced architectures, aims to enhance the accuracy and robustness of Alzheimer disease detection across the entire disease spectrum, ultimately contributing to improved patient care, a deeper understanding of Alzheimer disease, and the development of more effective treatments.

Review of Related Works

This section presents a detailed examination of prior research in Alzheimer's disease (AD) detection and classification, focusing on methodologies leveraging machine learning, deep learning, and transfer learning. Ullah and Jamjoom, (2023) utilized convolutional neural networks and Visual Geometry Group architecture for early Alzheimer's disease detection using magnetic resonance imaging scans. Their approach, which incorporated features extracted from the ImageNet dataset and the Alzheimer's Disease Neuroimaging Initiative dataset, achieved impressive accuracy rates of 99.27% for Alzheimer's disease vs. mild cognitive impairment and 97.06% for Alzheimer's disease vs. normal controls. Sethi et al., (2022) proposed a transfer learning framework utilizing the Efficient Net model, pre-trained on ImageNet, to classify individuals as Alzheimer's disease or cognitively normal. Their model achieved 91.36% accuracy and an area under the curve of 83%, underscoring the effectiveness of transfer learning in Alzheimer's disease diagnosis. Tanveer et al., (2022) presented Notably, the proposed ensemble outperformed snapshot ensembles and other existing deep models in similar studies conducted by other researchers. the Deep Transfer Ensemble (DTE) approach, which combines TL and hyperparameter randomization for enhanced diversity among neural networks. The method achieved 99.05% accuracy for differentiating CN from AD and 85% accuracy for small datasets, highlighting its robustness. Notably, the proposed ensemble outperformed snapshot ensembles and other existing deep models in similar studies conducted by other researchers. Danker and Wiklund, (2021) presented a novel approach that combined CNN and TL with data augmentation techniques, achieving a predictive accuracy of 97% for different AD stages. Despite using a smaller dataset, their model demonstrated significant effectiveness, with TL enabling knowledge transfer across stages. Sharma et al., (2022) proposed a modified inception model based on TL. The model incorporates normalization and data addition pre-processing techniques to enhance its performance. Remarkably, this proposed model achieved an impressive accuracy of 94.92% and a sensitivity of 94.94%. These results indicate that the proposed model surpassed other state-of-the-art models in AD detection. The training of the model utilizes a Kaggle dataset comprising 6,200 images, including 896 mild demented (M.D), 64 moderate demented (Mod.D), 3,200 non-demented (N.D), and 1,966 veritably mild demented (V.M.D) images. Agarwal et al., (2021) focused on segmenting and classifying AD using TL applied to MRI scans. Their customized CNN, pre-trained on segmented gray matter images, achieved a remarkable accuracy of 97.84%, validating its efficacy for early detection. Ghaffari et al., (2022) proposed TL-based CNN models, such as ResNet101, Xception, and InceptionV3, for AD classification using preprocessed and segmented MRI scans. Their models demonstrated superior performance, with InceptionV3 achieving 93.75% accuracy for binary and multiclass classification tasks. Kadri et al., (2022) investigated the application of the different CNN and transformers models on the early diagnosis of alzheimer. Further, we introduce a multi-modal method based on the MRI and PET modality for Alzheimer’s disease detection using the combination of the EfficientNetV2 and the vision transformer enhanced by a new data augmentation based on the self-attention generative adversarial networks (SAGAN). Their model, validated on ADNI and OASIS datasets, achieved an accuracy of 96%, showcasing the advantages of multimodal approaches. Blanton, (2023) proposed a novel 3D-to-2D conversion approach for brain imaging using a learnable weighted pooling (LWP) method, significantly reducing training time while achieving an accuracy of 88% for amyloid-beta PET imaging classification. Asgharzadeh-Bonab et al., (2023) explored feature- and decision-level fusion methods for dementia classification. Using EfficientNet-B7 with feature-level fusion, they achieved accuracies of 82.7% for CN-MCI and 89.7% for CN-AD scenarios, emphasizing the potential of fusion techniques in AD research. Saleh et al., (2023) highlighted the benefits of data augmentation and Dense Net architecture for handling high-dimensional MRI scans. Their model achieved an accuracy of 96.5% and an AUC of 99%, surpassing previous methods and showcasing the role of advanced augmentation techniques. Oommen & Arunnehru, (2023) utilized a three-step approach involving preprocessing, TL with autoencoders, and classification with a deep neural network (DNN). Their ensemble model achieved 98.54% accuracy, demonstrating its efficacy in AD stage classification. The method is compared to state-of-the-art approaches to validate its efficacy and performance. Agarwal et al., (2023) implemented EfficietNet-B0 convolutional neural network (CNN) with a novel approach "fusion of end-to-end and transfer learning" to classify different stages of AD. 245 T1W MRI scans of cognitively normal (CN) subjects, 229 scans of AD subjects, and 229 scans of subjects with stable mild cognitive impairment (sMCI) were employed achieving 95.29% accuracy for sMCI vs. AD classification and 87.38% for multiclass tasks, reinforcing CNN’s utility in AD analysis. Zheng et al., (2023) proposed a modified 3D Efficient Net architecture incorporating Mobile Inverted Bottleneck Convolution blocks. Their model demonstrated 95% accuracy for distinguishing NC from AD and 86.67% for NC vs. pMCI, validating its multiscale feature extraction capabilities. Mujahid et al., (2023) introduced an ensemble model combining EfficientNet-B2 and VGG-16 to address dataset imbalances. The ensemble achieved 97.35% accuracy and 99.64% AUC for multiclass datasets, highlighting its superior performance. The research reviewed underscores the transformative potential of deep learning and TL approaches for AD detection, with significant advancements achieved through innovative architectures, multimodal techniques, and comprehensive datasets. However, a notable research gap remains regarding the scalability and generalization. To address this gap, the proposed research aims to build upon Sharma et al. (2022) research by harnessing larger and more comprehensive datasets such as ADNI and OASIS. By combining these datasets and leveraging the NasNetMobile transfer learning architecture, which offers enhanced capacity for learning from large-scale data, our goal is to develop a model capable of accurate and efficient detection of the multiple stages of Alzheimer's disease.

MATERIALS AND METHOD

The paper aims to leverage transfer learning techniques to develop a model for detecting and classifying Alzheimer's disease based on medical images using Neural Architecture Search Network Mobile Architecture as depicted in Figure 3.1. The proposed methodology for the research on Alzheimer's disease detection using transfer learning with a Neural Architecture Search Network Mobile technique involves several key steps:

Proposed Methodology Architecture

Figure 3.1: Workflow of the Method

Figure 4.1 Training Session Output

4.4 Model evaluation

This section checks the performance of the model on the unseen data (that is, testing data), plot the graph of training accuracy and validation accuracy and plot the graph of training and validation loss for comparison shows the graph of training loss and validation loss for the first part of the training session. Loss is a measure of how well the model is performing the task it was trained for. Below is the accuracy and the loss of the said models. Figure 4.2 classification report of the model and the output of the testing. The test accuracy of 0.99% which is equal to 100% during the testing set was achieved.

Figure 4.2: Classification Report

Figure 4.3 shows the graph of training and validation accuracy for the first part training session. The accuracy was plotted against 100 epochs performed during the training session.

Figure 4.3 Training and Validation Accuracy

Figure 4.4: Training and Validation Loss plot

Figure 4.5: Confusion Matrix

Table 4.1 highlights significant differences in the performance of the benchmark models, the proposed approach. The study underscores the significance of dataset size in Alzheimer's disease detection using CNNs. By incorporating a larger dataset, our proposed approach achieved superior performance compared to the benchmark model. These findings emphasize the potential of NASNet Mobile architecture coupled with a comprehensive dataset for accurate and efficient AD detection.

CONCLUSION

This study highlights the critical role of dataset size and diversity in enhancing Alzheimer’s disease (AD) detection using deep learning models, particularly the NASNetMobile architecture. By integrating an expanded dataset and employing advanced transfer learning techniques, our proposed methodology achieved state-of-the-art accuracy, surpassing benchmark models. These results emphasize the transformative potential of leveraging robust datasets and sophisticated CNN architectures for accurate and efficient AD detection. Beyond its technical contributions, our work underscores the importance of data-driven innovation in advancing medical imaging and diagnostic methodologies. Moreover, this research provides valuable insights into the broader implications of CNN-based diagnostic approaches in healthcare, particularly for early disease detection and intervention. While our findings demonstrate significant progress, further validation on external datasets is essential to ensure generalizability and robustness across diverse populations and imaging protocols. Additionally, addressing ethical and societal considerations, such as patient privacy and equitable access to AI-driven technologies, remains a priority. Through continued research and collaboration, this work aims to pave the way for improved patient outcomes and the development of AI solutions that seamlessly integrate into clinical workflows, ultimately benefiting individuals affected by Alzheimer’s disease and their caregivers.                                        

REFERENCE

1. Agarwal, D., Berbís, M. Á., Luna, A., Lipari, V., Ballester, J. B., & de la Torre-Díez, I. (2023). Automated Medical Diagnosis of Alzheimer´s Disease Using an Efficient Net Convolutional Neural Network. Journal of Medical Systems, 47(1). https://doi.org/10.1007/s10916-023-01941-4
2. Agarwal, D., Marques, G., de la Torre-Díez, I., Franco Martin, M. A., García Zapiraín, B., & Martín Rodríguez, F. (2021). Transfer learning for alzheimer’s disease through neuroimaging biomarkers: A systematic review. Sensors, 21(21), 1–31. https://doi.org/10.3390/s21217259
3. Aging, N. I. on. (2003). alzheimer’ s disease Unraveling the Mystery. 1–65. https://www.bu.edu/alzresearch/files/pdf/adearunravelingtheMystery12-033.pdf
4. Asgharzadeh-Bonab, A., Kalbkhani, H., & Azarfardian, S. (2023). An Alzheimer’s disease classification method using fusion of features from brain Magnetic Resonance Image transforms and deep convolutional networks. Healthcare Analytics, 4(June), 100223. https://doi.org/10.1016/j.health.2023.100223
5. Ashraf, A., Naz, S., Shirazi, S. H., Razzak, I., & Parsad, M. (2021). Deep transfer learning for alzheimer neurological disorder detection. Multimedia Tools and Applications, 80(20), 30117–30142. https://doi.org/10.1007/s11042-020-10331-8
6. Bjorck, J., Gomes, C., Selman, B., & Weinberger, K. Q. (2018). Understanding batch normalization. Advances in Neural Information Processing Systems, 2018-Decem (NeurIPS), 7694–7705.
7. Blanton, H. (2023). Digital Commons @ Texas A & M University-San Antonio Efficient Training on Alzheimer’ s Disease Diagnosis with Learnable Weighted Pooling for 3D PET Brain Image Classification.
8. Chaihtra, D. H., & Vijaya Shetty, S. (2021). Alzheimer’s Disease Detection from Brain MRI Data using Deep Learning Techniques. 2021 2nd Global Conference for Advancement in Technology, GCAT 2021. https://doi.org/10.1109/GCAT52182.2021.9587756
9. Danker, A., & Wiklund, J. W. (2021). Using transfer learning to classify different stages of Alzheimer's disease [Master's thesis, KTH Royal Institute of Technology]. DiVA Academic Archive. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-297628
10. Ghaffari, H., Tavakoli, H., & Jahromi, G. P. (2022). Deep transfer learning–based fully automated detection and classification of Alzheimer’s disease on brain MRI. British Journal of Radiology, 95(1136). https://doi.org/10.1259/bjr.20211253
11. Ghazal, T. M., Abbas, S., Munir, S., Khan, M. A., Ahmad, M., Issa, G. F., Zahra, S. B., Khan, M. A., & Hasan, M. K. (2022). Alzheimer disease detection empowered with transfer learning. Computers, Materials and Continua, 70(3), 5005-5021. https://doi.org/10.32604/cmc.2022.020866
12. Ghosh, A., Sufian, A., Sultana, F., Chakrabarti, A., & De, D. (2019). Fundamental concepts of convolutional neural network. In Intelligent Systems Reference Library (Vol. 172, Issue January). https://doi.org/10.1007/978-3-030-32644-9_36
13. Ghosh, T., Palash, M. I. A., Yousuf, M. A., Hamid, M. A., Monowar, M. M., & Alassafi, M. O. (2023). A Robust Distributed Deep Learning Approach to Detect Alzheimer’s Disease from MRI Images. Mathematics, 11(12), 1–20. https://doi.org/10.3390/math11122633
14. Gupta, N., & Agrawal, R. (2021). Impact of Deep Learning on Semantic Sentiment Analysis. February, 97–117. https://doi.org/10.4018/978-1-7998-7511-6.ch007
15. Hao, W., Yizhou, W., Yaqin, L., & Zhili, S. (2020). The Role of Activation Function in CNN. Proceedings - 2020 2nd International Conference on Information Technology and Computer Application, ITCA 2020, 429–432. https://doi.org/10.1109/ITCA52113.2020.00096
16. Hunter, J. D., Dale, D., Firing, E., Droettboom, M., & the Matplotlib development team. (2017). Matplotlib: Python plotting library (Version 2.0.2) [Computer software]. https://matplotlib.org/
17. Iorkaa, A. A. (2021). Machine Learning Techniques, methods and Algorithms: Conceptual and Machine Learning Techniques, methods and Algorithms: Conceptual and Practical Insights. 11(August), 55–64. https://doi.org/10.9790/9622-1108025564
18. Johnson, K. A., Fox, N. C., Sperling, R. A., & Klunk, W. E. (2012). Brain imaging in Alzheimer disease. Cold Spring Harbor Perspectives in Medicine, 2(4). https://doi.org/10.1101/cshperspect.a006213
19. Kadri, R., Bouaziz, B., Tmar, M., & Gargouri, F. (2022). Multimodal deep learning based on the combination of EfficientNetV2 and ViT for Alzheimer’s disease early diagnosis enhanced by SAGAN data augmentation. International Journal of Computer Information Systems and Industrial Management Applications, 14, 313–325.
20. Kaggle. (2019). Alzheimer MRI Preprocessed Dataset. https://www.kaggle.com/datasets/sachinkumar413/alzheimer-mri-dataset
21. Waskom, M. L. (2021). seaborn: Statistical data visualization. Journal of Open Source Software, 6(60), 3021. https://doi.org/10.21105/joss.03021
22. ?awrynowicz, A., & Tresp, V. (2014). Introducing machine learning. Perspectives on Ontology Learning, 18(January), 35–50. https://doi.org/10.1007/978-3-030-67626-1_8
23. Liu, Q., & Wu, Y. (2012). Encyclopedia of the Sciences of Learning. Encyclopedia of the Sciences of Learning, January 2012. https://doi.org/10.1007/978-1-4419-1428-6
24. Mathew, A., Amudha, P., & Sivakumari, S. (2021). Deep learning techniques: an overview. Advances in Intelligent Systems and Computing, 1141(January), 599–608. https://doi.org/10.1007/978-981-15-3383-9_54
25. McKinney, W., & Team, P. D. (2015). Pandas - Powerful Python Data Analysis Toolkit. Pandas - Powerful Python Data Analysis Toolkit, 1625.
26. Miceli, P. A., Blair, W. D., & Brown, M. M. (2018). Isolating Random and Bias Covariances in Tracks. In 2018 21st International Conference on Information Fusion, FUSION 2018. https://doi.org/10.23919/ICIF.2018.8455530
27. Mujahid, M., Rehman, A., Alam, T., Alamri, F. S., Fati, S. M., & Saba, T. (2023). An Efficient Ensemble Approach for Alzheimer’s Disease Detection Using an Adaptive Synthetic Technique and Deep Learning. Diagnostics, 13(15), 1–20. https://doi.org/10.3390/diagnostics13152489
28. Naik, P. G. (2021). Conceptualizing Python in Google COLAB Hands-on Practical Sessions (Issue January). https://www.google.co.id/books/edition/Conceptualizing_Python_in_Google_COLAB/
29. Name, S., Chugh, L., Id, S., Name, S., Alshabrawy, O., Supervisor, S., & Science, D. (n.d.). MSc Computer Science and Digital Technologies Project.
30. Naz, S., Ashraf, A., & Zaib, A. (2022). Transfer learning using freeze features for Alzheimer neurological disorder detection using ADNI dataset. Multimedia Systems, 28(1), 85–94. https://doi.org/10.1007/s00530-021-00797-3
31. Nirthika, R., Manivannan, S., Ramanan, A., & Wang, R. (2022). Pooling in convolutional neural networks for medical image analysis: a survey and an empirical study. Neural Computing and Applications, 34(7), 5321–5347. https://doi.org/10.1007/s00521-022-06953-8
32. Oommen, D. K., & Arunnehru, J. (2023). Alzheimer’s Disease Stage Classification Using a Deep Transfer Learning and Sparse Auto Encoder Method. Computers, Materials and Continua, 76(1), 793–811. https://doi.org/10.32604/cmc.2023.038640
33. Ramani, A., Jensen, J. H., & Helpern, J. A. (2006). Quantitative MR imaging in Alzheimer disease. Radiology. https://doi.org/10.1148/radiol.2411050628
34. Saleh, A. W., Gupta, G., Khan, S. B., Alkhaldi, N. A., & Verma, A. (2023). An Alzheimer’s disease classification model using transfer learning Densenet with embedded healthcare decision support system. Decision Analytics Journal, 9(October), 100348. https://doi.org/10.1016/j.dajour.2023.100348
35. Sarasaen, C., Chatterjee, S., Breitkopf, M., Rose, G., Nürnberger, A., & Speck, O. (2021). Fine-tuning deep learning model parameters for improved super-resolution of dynamic MRI with prior-knowledge. Artificial Intelligence in Medicine, 121. https://doi.org/10.1016/j.artmed.2021.102196
36. Saxen, F., Werner, P., Handrich, S., Othman, E., Dinges, L., & Al-hamadi, A. (2019). Face Attribute Detection with MobileNetV2 and. October. https://doi.org/10.1109/ISPA.2019.8868585
37. Sethi, M., Ahuja, S., Singh, S., Snehi, J., & Chawla, M. (2022). An Intelligent Framework for Alzheimer’s disease Classification Using EfficientNet Transfer Learning Model. 2022 International Conference on Emerging Smart Computing and Informatics, ESCI 2022, June. https://doi.org/10.1109/ESCI53509.2022.9758195
38. Sharma, C., & Parikh, S. (2022). Transfer Learning and its application in Computer Vision: A Review. March. https://arxiv.org/abs/
39. Sharma, S., Gupta, S., Gupta, D., Juneja, S., Mahmoud, A., El–Sappagh, S., & Kwak, K. S. (2022). Transfer learning-based modified inception model for the diagnosis of Alzheimer’s disease. Frontiers in Computational Neuroscience, 16. https://doi.org/10.3389/fncom.2022.1000435
40. Shen, Q., Loewenstein, D. A., Potter, E., Zhao, W., Appel, J., Greig, M. T., Raj, A., Acevedo, A., Schofield, E., Barker, W., Wu, Y., Potter, H., & Duara, R. (2011). Volumetric and visual rating of magnetic resonance imaging scans in the diagnosis of amnestic mild cognitive impairment and Alzheimer’s disease. Alzheimer’s and Dementia, 7(4), e101–e108. https://doi.org/10.1016/j.jalz.2010.07.002
41. States, U., & Alzheimer, A. (1906). Alzheimer’ s Disease. Alzheimer, A. (1906/1987). About a peculiar disease of the cerebral cortex. Alzheimer Disease and Associated Disorders, 1(1), 3-8. (Original work published 1906)
42. Tanveer, M., Rashid, A. H., Ganaie, M. A., Reza, M., Razzak, I., & Hua, K. L. (2022). Classification of Alzheimer’s Disease Using Ensemble of Deep Neural Networks Trained Through Transfer Learning. IEEE Journal of Biomedical and Health Informatics, 26(4), 1453–1463. https://doi.org/10.1109/JBHI.2021.3083274
43. Thakur, A. (2021). Fundamentals of Neural Networks. International Journal for Research in Applied Science and Engineering Technology, 9(VIII), 407–426. https://doi.org/10.22214/ijraset.2021.37362
44. Ting, K. M. (2017). Confusion Matrix. Encyclopedia of Machine Learning and Data Mining, October, 260–260. https://doi.org/10.1007/978-1-4899-7687-1_50
45. TÜRK, F., & KÖKVER, Y. (2022). Application with deep learning models for COVID-19 diagnosis. Sakarya University Journal of Computer and Information Sciences, 5(2), 169–180. https://doi.org/10.35377/saucis...1085625
46. Ullah, Z., & Jamjoom, M. (2023). A Deep Learning for Alzheimer’s Stages Detection Using Brain Images. Computers, Materials and Continua, 74(1), 1457–1473. https://doi.org/10.32604/cmc.2023.032752
47. Wang, H., & Raj, B. (2017). On the Origin of Deep Learning. February.
48. Wright, A. (2010). Brain Scanning Techniques (CT, MRI, fMRI, PET, SPECT, DTI, DOT). Cerebra Positively Different, 1089812, 1–14.
49. Wu, H., & Gu, X. (2015). Towards dropout training for convolutional neural networks. Neural Networks, 71, 1–10. https://doi.org/10.1016/j.neunet.2015.07.007
50. Yang, F., Zhang, W., Tao, L., & Ma, J. (2020). Transfer learning strategies for deep learning-based PHM algorithms. Applied Sciences (Switzerland), 10(7). https://doi.org/10.3390/app10072361
51. Yang, H. D., Kim, D. H., Lee, S. B., & Young, L. D. (2016). History of Alzheimer’s Disease. Dementia and Neurocognitive Disorders, 15(4), 115. https://doi.org/10.12779/dnd.2016.15.4.115
52. Zheng, B., Gao, A., Huang, X., Li, Y., Liang, D., & Long, X. (2023). A modified 3D EfficientNet for the classification of Alzheimer’s disease using structural magnetic resonance images. IET Image Processing, 17(1), 77–87. https://doi.org/10.1049/ipr2.12618
53. Zoph, B., Vasudevan, V., Shlens, J., & Le, Q. V. (2018). Learning Transferable Architectures for Scalable Image Recognition. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 8697–8710. https://doi.org/10.1109/CVPR.2018.00907.