AI-Driven Alzheimer’s Detection Performance Analysis of Pretrained CNN Models on MRI Data

Abstract

Alzheimer\'s disease (AD) is a memory and cognitive function neurodegenerative disease. Early diagnosis is important for appropriate treatment, but conventional diagnosing techniques are not very effective. In this paper, automatic detection of Alzheimer\'s disease from MRI scans using deep learning techniques is discussed. We trained and compared six pre-trained Convolutional Neural Network (CNN) models—VGG19, ResNet101, EfficientNetB3, MobileNetV2, InceptionV3, and DenseNet121—to differentiate MRI scans as Alzheimer and Non-Alzheimer. In experiments, InceptionV3 is the optimal model for classification of Alzheimer\'s disease and all others are inferior with an accuracy of 99.2%. Confusion matrices, classification reports, and accuracy metrics were utilized to test performance. We also created a web-based real-time MRI classification tool with a user-friendly interface for clinicians. This work showcases the potential of deep learning in medical imaging and the detection of early disease, and opens the door to AI-augmented improvement in Alzheimer\'s diagnosis.

Introduction

Millions suffer from Alzheimer's disease (AD), a neurodegenerative disorder causing memory loss and cognitive decline. Early and accurate diagnosis is crucial but challenging with traditional mental and physical tests, which often miss early stages. Recent advances in deep learning and medical imaging have enabled automated, precise AD diagnosis from MRI scans.

This study compares six pre-trained convolutional neural network (CNN) models—VGG19, ResNet101, EfficientNetB3, MobileNetV2, InceptionV3, and DenseNet121—using a publicly available MRI dataset from Kaggle to classify brains as Alzheimer’s or non-Alzheimer’s. Among these, InceptionV3 achieved the highest accuracy at 99.2%. A web application was also developed to allow real-time diagnosis by uploading MRI scans.

The paper highlights deep learning’s potential in medical imaging to aid early AD diagnosis, improving healthcare outcomes and supporting AI-driven neurodegenerative disease detection.

The literature review outlines prior research on machine learning and deep learning approaches for brain disease detection using MRI data, emphasizing model optimization, feature extraction, and classification techniques.

The methodology involves dataset preprocessing, training with TensorFlow and Keras frameworks, and model evaluation with metrics provided by Scikit-Learn. Key CNN layers used include Conv2D, Flatten, MaxPooling, AveragePooling, Dense, Activation, and Dropout, all contributing to effective feature extraction and classification.

Detailed descriptions of models are provided: VGG16 achieved moderate accuracy (~84%), while ResNet101 performed better (~98.3%) due to its deep architecture and skip connections that mitigate vanishing gradient issues, enabling identification of complex Alzheimer’s features.

Overall, the study demonstrates that deep learning models, particularly InceptionV3, can significantly enhance early Alzheimer’s diagnosis using MRI scans, paving the way for AI-assisted medical diagnostics.

Conclusion

We learned how to use deep learning to forecast Alzheimer\'s disease from MRI scans in this article. We compared six pre-trained Convolutional Neural Network ( CNN) models—VGG19, ResNet101, EfficientNetB3, MobileNetV2, InceptionV3, and DenseNet121—to find out how much classification ability they have. Models were able to learn the disease patterns due to the dataset, which was gathered from Kaggle and contained MRI images that were Alzheimer\'s and non-Alzheimer. According to our findings, InceptionV3 was the highest-scoring model with a peak accuracy of 99.2 % for the detection of Alzheimer\'s disease, with DenseNet121 performing the lowest at 38.5%, accepting how inefficient it was for this instance, with others like ResNet101 (98.3%) and EfficientNetB3 (93.6%) also doing well. Some of the performance metrics that ensured the credibility of the top-scoring model included confusion matrix, accuracy plot, loss plot, and classification report. We created an online platform for real-time Alzheimer\'s diagnosis to make it available, where the patient can submit MRI scans to be diagnosed in real-time. Medical practitioners can use this application to assist in the detection of early, which will result in early interventions and better patient outcomes. Additional expansion of the dataset with more sophisticated deep learning models and investigation of explainable AI methods to increase interpretability are all promising directions for future studies. Using AI methods, our study demonstrates the potential of deep learning in medical imaging and its ability to improve the detection of Alzheimer\'s disease.

References

[1] Khan P, Kader MF, Islam SR, Rahman AB, Kamal MS, Toha MU, et al.. Machine learning and deep learning approaches for brain disease diagnosis: principles and recent advances. IEEE Access.. [2] Prajapati R, Khatri U, Kwon GR. “An efficient deep neural network binary classifier for alzheimer\'s disease classification,” In: International Conference on Artificial Intelligence in Information and Communication (ICAIIC).. [3] Helaly HA, Badawy M, Haikal AY. Deep learning approach for early detection of Alzheimer\'s disease. Cogn Computing. [4] Martinez-Murcia FJ, Ortiz A, Gorriz JM, Ramirez J, Castillo-Barnes D. Studying the manifold structure of Alzheimer\'s disease: a deep learning approach using convolutional autoencoders. IEEE J Biomed Health Inform. (2020). [5] Ronghui Ju , Chenhui Hu, Pan Zhou , and Quanzheng Li, “Early Diagnosis of Alzheimer’s Disease Based on Resting-State Brain Networks and Deep Learning”. [6] Ruoxuan Cuia, Manhua Liu “RNN-based longitudinal analysis for diagnosis of Alzheimer’s disease”, Informatics in Medicine Unlocked, 2019. [7] Ye, D.H., Pohl, K.M., Davatzikos, C., 2011. Semi-supervised pattern classification: application to structural MRI of Alzheimer’s disease. Pattern Recognition in NeuroImaging(PRNI),International Workshop on. IEEE,(2011). [8] Filipovych, R., Davatzikos, C., Semi-supervised pattern classification of medical images: application to mild cognitive impairment (MCI). [9] kaggle. Alzheimer\'s Dataset [Online]. Available on kaggle url https://www.kaggle.com/datasets/tourist55/alzheimers-dataset-4-class -of-images.

Copyright

Copyright © 2025 Varun Patel , Priyank Mishra , Priyanshu Rawat , Yash Upadhyay , Neha Prakash . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.