MindCareAI-Alzheimer’s Disease Detection from Scan Using Machine Learning with Chatbot and Hospital Integration

MindCareAI-Alzheimers Disease Detection from Scan Using Machine Learning with Chatbot and Hospital Integration

Atul R. Bodke (1), Nitin D. Palve (2), Sudeep S. Avhad (3), Mayur S. Aher (4) and Prof. Priyanka Narode (5)

SND College of Engineering and Research Centre, Yeola, Nashik, Maharashtra, India 423401

Abstract – Alzheimer’s disease is an unrepairable degenerative brain disease. Every four seconds, someone in the world is diagnosed with Alzheimer’s disease. The result is fatal, as it leads to death. As a result, it’s crucial to catch the disease early on. The leading cause of dementia is Alzheimer’s disease. Dementia causes a reduction in reasoning abilities and interpersonal coping skills, which affects people’s ability to function independently. The patient will forget recent events in the early stages. If the illness progresses, they will gradually forget whole events. It is essential to diagnose the disease as soon as possible. This paper proposes a model that takes brain MRI sample images as input and determines whether a person has mild, moderate, or no Alzheimer’s disease as an output. We are using the VGG19 and DenseNet169 architectures for this classification, providing a comparative analysis of which architecture shows promising results.

Keywords – Alzheimer’s, MRI Images, VGG19, CNN Densenet

With the rapid advancement of Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL), intelligent healthcare systems are becoming more effective in medical diagnosis and disease prediction. Machine learning algorithms can analyze complex MRI brain scan patterns, identify early biomarkers of Alzheimers Disease, and provide predictive insights with high accuracy. Deep learning techniques, especially Convolutional Neural Networks (CNNs), have shown significant performance in medical image analysis by automatically extracting important features from MRI images without manual intervention.

The integration of AI with medical imaging helps radiologists and neurologists by providing faster, more consistent, and data-driven diagnostic support. AI-based systems can reduce diagnostic errors, improve prediction accuracy, and assist healthcare professionals in making timely clinical decisions. Furthermore, explainable AI techniques such as Grad-CAM visualization improve transparency by highlighting affected brain regions responsible for the prediction results.

1. INTRODUCTION

   Alzheimers Disease (AD) is a progressive neurodegenerative disorder that primarily affects memory, thinking, learning ability, and behavior. It is one of the leading causes of dementia worldwide, affecting millions of people every year, especially older adults. As the disease progresses, patients experience memory loss, confusion, difficulty in communication, and a decline in daily functional activities. Alzheimers not only impacts patients physically and mentally but also creates emotional and financial challenges for families and healthcare systems.

   Early and accurate detection of Alzheimers Disease is extremely important because timely diagnosis can help in better patient management, treatment planning, and slowing down disease progression. Traditional diagnostic methods mainly depend on manual interpretation of MRI brain scans, cognitive assessments, and clinical examinations performed by medical experts. However, these methods are often time-consuming, costly, and prone to human error, particularly in the early stages of the disease where symptoms may be difficult to identify.

   To address these challenges, the proposed system, MindCareAI, is developed as an intelligent web-based healthcare platform for Alzheimers Disease detection and patient assistance. The system uses MRI brain scans and a CNN-based deep learning model to predict the stage of Alzheimers Disease accurately. In addition to disease prediction, the application also generates automated diagnostic reports and recommends nearby doctors or hospitals for further consultation and treatment.

   The developed web application provides a complete AI-driven healthcare solution by integrating disease detection, explainable AI visualization, automated report generation, and healthcare connectivity within a single platform. The main objective of MindCareAI is to improve diagnostic efficiency, accessibility, and patient-centered healthcare support for Alzheimers Disease management.

   Fig. 1. Image representing a Healthy Brain vs. Severe AD Brain

   The proposed system, MindCareAI, combines machine learningbased Alzheimers detection with a patient-interactive chatbot and hospital integration module. The system allows users to upload MRI scans, which are processed through a trained ML model to detect potential Alzheimers indicators. The built-in chatbot enables users to interact with the system, receive disease predictions in an understandable format, and ask health-related questions for better awareness. Additionally, the platform is integrated with hospital and doctor databases to display nearby medical centers and specialists for immediate consultation.

2. LITERATURE SURVEY

   Alzheimers Disease (AD) is one of the most common neurodegenerative disorders affecting memory, thinking ability, and cognitive functions. With the advancement of Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL), several researchers have developed automated systems for early Alzheimers detection using MRI brain images. Existing studies mainly focus on improving prediction accuracy, interpretability, and computational efficiency.

   Sharma et al. [1] proposed a Deep Transfer Learning-Based Automated Alzheimers Detection system using the VGG16 architecture. Their approach utilized MRI brain images along with Grad-CAM visualization to identify affected regions of the brain. The model achieved high accuracy in detecting Alzheimers disease and improved diagnostic interpretation. However, the system required large training datasets and lacked integration with real-time healthcare support services.

   Lee et al. [2] introduced an Explainable Alzheimers Diagnosis framework using Vision Transformers (ViT) and Grad-CAM techniques. Their study focused on enhancing transparency and interpretability in AI-based diagnosis to increase clinician trust. Although the framework provided better visualization and prediction performance, it demanded high computational resources and GPU support, making deployment difficult in resource-constrained environments.

   Ahmed et al. [3] developed an Ensemble CNN model combining ResNet50 and VGG19 architectures for Alzheimers stage classification. Their approach improved classification accuracy for mild, moderate, and severe stages of Alzheimers

   disease. Despite achieving better precision, the model increased computational complexity and was less suitable for real-time clinical applications.

   Kumar and Patel [4] proposed a Lightweight CNN model designed for edge devices and portable healthcare systems. Their method reduced inference time while maintaining acceptable accuracy for Alzheimers detection. However, the model performance decreased when handling noisy or low-quality MRI images, indicating the need for robust preprocessing techniques.

   Chen et al. [5] explored multimodal deep learning by integrating MRI data with genetic information using Deep CNN architectures. Their study demonstrated improved prediction performance through the combination of multiple healthcare data sources. However, the approach faced challenges due to the limited availability of multimodal dataset and lacked practical web-based clinical integration.

   Verma et al. [6] implemented Grad-CAM visualization techniques for interpretable Alzheimers diagnosis using CNN-based models. Their work enhanced explainability and improved medical acceptance of AI systems by visually highlighting affected brain regions. However, the framework was limited to specific CNN layers and did not provide additional healthcare assistance functionalities.

   From the reviewed literature, it is observed that most existing systems primarily focus on disease classification and prediction accuracy. Very few studies provide an integrated healthcare solution that combines disease detection, automated report generation, and doctor recommendation services within a single platform.

   The proposed system, MindCareAI, aims to overcome these limitations by developing an intelligent web-based healthcare application for Alzheimers detection using MRI images. The system integrates machine learning-based disease prediction, automated diagnostic report generation, and nearby doctor or hospital recommendation features. This approach improves accessibility, user interaction, and healthcare support, making the diagnosis process more efficient and user-friendly.

3. PROPOSED WORK

   The Proposed System, Mindcareai, Is An Intelligent Web-Based Healthcare Platform Developed For The Early Detection And Analysis Of Alzheimers Disease (Ad) Using Mri Brain Images And Deep Learning Techniques. The System Aims To Automate Disease Prediction While Also Providing Healthcare Support Services Such As Report Generation And Nearby Doctor

   Recommendations. By Integrating Artificial Intelligence (Ai), Medical Image Processing, And Healthcare Connectivity, The Proposed System Enhances Diagnostic Accuracy, Accessibility, And Patient Support Within A Single Platform. The System Mainly Consists Of Three Integrated Modules: The Alzheimers Disease Detection Module, Automated Report Generation Module, And Doctor Recommendation Module. The Workflow Begins When The User Uploads An Mri Brain Scan Through The Web Application. The Uploaded Image Undergoes

   Preprocessing Operations Such As Resizing, Normalization, And Noise Removal To Improve Image Quality And Maintain Consistency In The Dataset. These Preprocessing Steps Help The Deep Learning Model Extract Meaningful Features More Effectively.

   After Preprocessing, The Mri Image Is Analyzed Using A Convolutional Neural Network (Cnn)-Based Deep Learning Model Trained On Alzheimers Mri Datasets. The Model Automatically Extracts Spatial And Structural Brain Features To Identify Abnormalities Associated With Alzheimers Disease.

   Based On The Extracted Features, The System Classifies The Mri Scan Into Different Disease Categories Such As Non-Demented, Very Mild Demented, Mild Demented, And Moderate Demented.

   Once The Prediction Process Is Completed, The System Generates An Automated Diagnostic Report Containing The Predicted Disease Stage, Confidence Score, And Analysis Summary. The Report Is Generated In Pdf Format And Can Be Downloaded Or Shared With Healthcare Professionals For Further Medical Evaluation. To Improve Transparency And Explainability, Grad-Cam Visualization Techniques Are Used To Highlight The Affected Brain Regions Responsible For The Prediction.

   In Addition To Prediction And Report Generation, The Proposed System Includes A Doctor Recommendation Module That Helps Users Locate Nearby Neurologists Or Hospitals For Consultation. Based On The Diagnosis Results, The Web Application Provides Healthcare Suggestions And Connects Patients With Medical Specialists For Further Treatment And Guidance. This Feature Improves Accessibility To Medical Services And Supports Timely Clinical Intervention. The Proposed Mindcareai System Provides A Complete Aidriven Healthcare Solution By Combining Deep Learning-Based Disease Prediction, Automated Medical Reporting, Explainable Ai Visualization, And Healthcare Connectivity. The Platform Is Designed To Be User-Friendly, Efficient, And Accessible, Making Alzheimers Diagnosis And Consultation Support More Effective For Patients And Healthcare Providers.

   A. Dataset

   The Dataset Used In This Project Is Obtained From The Opensource Kaggle Alzheimer Mri Dataset. The Dataset Contains Approximately 6,000 Mri Brain Scan Images Categorized Into Four Classes: Non-Demented, Very Mild Demented, Mild Demented, And Moderate Demented. These Categories

   Represent Different Stages Of Alzheimers Disease Progression. The Dataset Is Divided Into Training And Testing Sets Using An 80:20 Ratio, Where 80% Of The Data Is Used For Model Training And 20% Is Used For Validation And Testing.

   Maintaining The Same Distribution Across Training And Testing Datasets Ensures Fair Model Evaluation And Improves Prediction Consistency.

   Before Training, The Mri Images Undergo Preprocessing Techniques Such As Image Resizing, Normalization, And Noise Reduction. These Preprocessing Operations Improve Image Clarity And Enhance The Performance Of The Deep Learning Model During Feature Extraction And Classification.

   Fig. 2. Dataset After Pre-Processing B. Methodology

   The Methodology Of Mindcareai Follows A Systematic Workflow For Alzheimers Disease Detection And Healthcare Integration. The Process Starts When The Patient Uploads An Mri Brain Scan Through The Web Application Interface. The Uploaded Image Is Stored Securely In The Patient Management Module.

   The Mri Scan Is Then Passed To The Disease Prediction Module, Where A Trained Cnn-Based Deep Learning Model Analyzes The Image To Detect Alzheimers Disease And Determine Its Stage. The Model Extracts Important Image Features Automatically And Classifies The Disease Stage Accurately.

   To Improve Explainability And User Trust, Grad-Cam Visualization Techniques Are Applied To Highlight The Affected Brain Regions Responsible For The Prediction. Based On The Prediction Results, The System Automatically Generates A Detailed Pdf Diagnostic Report Containing Disease Stage Information, Confidence Score, And Visualization Outputs.

   After Report Generation, The Doctor Recommendation Module Identifies Nearby Neurologists And Hospitals To Assist Patients In Seeking Professional Medical Consultation. The System Finally Connects Patients With Healthcare Providers

   Through The Doctor Management System For Further Diagnosis And Treatment Support.

   Fig. 3. Proposed Methodology C. Convolutional Neural Network For Alzheimers Disease Detection

   In The Proposed System, A Convolutional Neural Network (Cnn) Is Used For Alzheimers Disease Classification Using Mri Brain Images. Cnn Is A Powerful Deep Learning (Dl) Technique Widely Used For Medical Image Analysis Due To Its Ability To Automatically Learn Important Visual Features.

   The Cnn Architecture Consists Of Multiple Convolutional Layers, Pooling Layers, Activation Functions, And Fully Connected Dense Layers. The Convolutional Layers Extract Important Spatial Features From Mri Images Using Different Kernels To Identify Alzheimers-Affected Brain Regions. The Rectified Linear Unit (Relu) Activation Function Is Applied To Introduce Non-Linearity And Improve Learning Performance.

   Pooling Layers, Especially Max Pooling Layers, Are Used To Reduce Image Dimensions And Computational Complexity While Preserving Important Features. These Layers Help Eliminate Unnecessary Image Regions That Are Not Related To

   Alzheimers Disease.

   Finally, The Extracted Features Are Passed Through Fully Connected Dense Layers, Where The Learning And Classification Process Takes Place. The Cnn Model Classifies Mri Images Into Different Alzheimers Stages Such As Non-Demented, Very Mild Demented, Mild Demented, And Moderate Demented Based On The Learned Patterns And Extracted Features.

4. CONCLUSION

   The proposed system, MindCareAI, successfully demonstrates the integration of Artificial Intelligence, deep learning, medical imaging, and web tecnologies for the early detection of Alzheimers Disease (AD). The developed web application provides an intelligent and user-friendly platform capable of analyzing MRI brain scans and accurately predicting the stage of Alzheimers disease using a Convolutional Neural Network (CNN)-based deep learning model.The system effectively

   performs MRI image preprocessing, disease stage prediction, automated report generation, and visualization of affected brain regions using Grad-CAM techniques. The generated diagnostic report helps users and healthcare professionals better understand the prediction results and supports further medical evaluation. In addition to disease detection, the application also includes a Doctor Recommendation Module that assists patients in locating nearby neurologists and hospitals for timely consultation and treatment. This integration improves healthcare accessibility and provides a complete patient-centered solution within a single platform.The developed web application offers several advantages such as improved diagnostic accuracy, reduced manual effort, faster prediction, explainable AI-based visualization, and enhanced patient interaction. By combining disease prediction, automated medical reporting, and healthcare connectivity, MindCareAI represents an effective and practical solution for intelligent Alzheimers Disease diagnosis and management.Overall, the project demonstrates how AI-driven healthcare systems can support early diagnosis and improve medical decision-making. In the future, the system can be further enhanced by integrating real-time clinical datasets, advanced deep learning models, multilingual chatbot support, and cloud-based healthcare services to provide even more accurate and scalable healthcare solutions.

5. REFERENCES

1. Sharma, R., Mehta, S., And Gupta, A., Deep Transfer Learning-Based Automated Alzheimers Detection Using Mri Images, Proceedings Of The International

   Conference On Medical Imaging And Deep Learning, 2023.

2. Lee, J., Kim, H., And Park, S., Explainable Alzheimers

   Diagnosis Using Vision Transformers And Gradcam, Ieee Transactions On Neural Systems And Rehabilitation Engineering, 2024.

3. Ahmed, T., Khan, M., And Yousaf, M., Ensemble Cnn

   For Alzheimers Stage Classification, Journal Of Computational Neuroscience, 2022.

4. Kumar, P., And Patel, D., Lightweight Cnn For Alzheimers Detection On Edge Devices, International Journal Of Artificial Intelligence And Applications, 2023.

5. Chen, L., Wang, Z., And Zhao, Y., Alzheimers Disease Detection Using Multimodal Data And Deep

   Learning, Frontiers In Computational Neuroscience, 2023.

6. Verma, R., Singh, N., And Bhatia, K., Grad-Cam Visualization For Interpretable Alzheimers Diagnosis, Ieee Access, 2022.