Department of Information Technology

DAV PROJECT REPORT

ON
Alzheimer Detection System

Submitted by:
Himanshu Jain (2200911540054)
Divyansh Arya (2200911540043)
Ashish Gupta (2200911540032)

JSS Academy of Technical Education, NOIDA

Dr. APJ Abdul Kalam Technical University, Lucknow, U.P

Session 2024-25
 DECLARATION

We hereby declare that this submission is our own work and

that, to the best of my knowledge and belief, it contains no
material previously published or written by another person nor
material which to a substantial extent has been accepted for the
award of any other degree or diploma of the university or other
institute of higher learning, except where due acknowledgment
has been made in the text.
 CERTIFICATE

This is to certify that Project Report entitled “Alzheimer’s Detection

System” which is submitted by Us in partial fulfillment of the
requirement for the award of degree B. Tech. in Department of
Information Technology of U. P. Technical University, is a record
of the candidate own work carried out by her under my supervision.
The matter embodied in this thesis is original and has not been
submitted for the award of any other degree.

Supervisor:
Mrs. Aparna Shrivastava
(Associate Professor Dept. of IT)
Introduction
Alzheimer's disease is a progressive neurological disorder that affects
millions of people worldwide. Early detection is crucial for effective
treatment and management of the disease. In this project, we
developed a machine learning-based system that predicts the
likelihood of Alzheimer's by analysing uploaded brain scan images.
The application integrates advanced machine learning algorithms with
a user-friendly web interface, enabling healthcare professionals and
researchers to make informed decisions. The project's objective is to
leverage technology to enhance diagnostic accuracy and accessibility.
Technologies Used
The project employs a robust set of technologies and frameworks to
achieve its goals:
1. Programming Language: Python was chosen due to its vast
library ecosystem and ease of integration with machine learning
frameworks.
2. Libraries and Frameworks:
o Numpy: For efficient numerical computations and matrix
operations.
o Pandas: For data manipulation and preprocessing.
o Matplotlib: To visualize data distributions and model
performance.
o TensorFlow and Keras: For designing, training, and
optimizing deep learning models.
3. Web Development:
 o Flask: Used as both the backend and frontend framework,
facilitating seamless user interaction with the machine
learning model. It supports functionalities such as image
upload, navigation between pages, and result display.
Methodology
Data Collection
The model was trained using a dataset of brain scan images that
included labeled examples of healthy and Alzheimer's-affected scans.
This dataset was preprocessed to ensure high-quality inputs for the
machine learning model.
Preprocessing
Preprocessing is a critical step in machine learning to ensure the data
is clean, normalized, and ready for training. The preprocessing steps
included:
• Image Resizing: All brain scan images were resized to a
uniform dimension to ensure consistency.
• Normalization: Pixel values were normalized to bring all data
to the same scale, aiding in faster convergence during training.
• Data Augmentation: Techniques such as rotation, flipping, and
zooming were applied to increase the dataset size and improve
model robustness.
Model Design and Training
A Convolutional Neural Network (CNN) architecture was used due to
its exceptional performance in image classification tasks. The key
components of the model design include:
1. Input Layer: Takes the preprocessed images as input.
2. Convolutional Layers: Extract spatial features from the images.
3. Pooling Layers: Reduce dimensionality while retaining
essential features.
 4. Fully Connected Layers: Combine extracted features to make
predictions.
5. Output Layer: A softmax activation function outputs the
probabilities of Alzheimer's presence.
The model was trained using TensorFlow and Keras frameworks.
Binary cross-entropy loss was chosen as the loss function, and the
Adam optimizer was used for faster convergence.
Hyperparameters:
• Batch Size: 32
• Learning Rate: 0.001
• Epochs: 50
Evaluation
To evaluate the model's performance, metrics such as accuracy,
precision, recall, and F1-score were calculated. The model achieved a
high prediction ratio, indicating its effectiveness in distinguishing
between healthy and Alzheimer's-affected scans.
Web Application Development
The web application serves as the interface for users to interact with
the machine learning model. Flask was used to develop a lightweight
and responsive application.
Features of the Web Application
1. Home Page: Provides an overview of the application and its
purpose.
2. Upload Page: Allows users to upload brain scan images for
analysis.
3. Results Page: Displays the prediction results along with
probabilities.
4. About Page: Explains the technology stack and the project's
objectives.
 5. References Page: Lists the resources and research materials
used in the project.
The Flask backend handles:
• Image uploads and validation.
• Model inference to generate predictions.
• Routing between various pages.
The frontend integrates HTML, CSS, and JavaScript to ensure a user-
friendly experience.
Challenges Faced
Dataset Limitations
One major challenge was obtaining a sufficiently large and diverse
dataset for training. The quality and variability of the brain scan
images significantly impact the model's accuracy.
Computational Resources
Training a deep learning model requires substantial computational
power. To address this, GPU-accelerated environments were used for
training.
Overfitting
Initially, the model showed signs of overfitting, performing well on
the training data but poorly on validation data. This was mitigated by:
• Implementing dropout layers in the CNN architecture.
• Using data augmentation techniques.
• Early stopping during training.
Integration of Frontend and Backend
Ensuring smooth communication between the Flask backend and the
frontend required meticulous debugging and testing.
Results
The project successfully developed a system capable of predicting
Alzheimer's with high accuracy. The model was able to:
• Achieve an accuracy of 92% on the test set.
• Demonstrate a precision of 0.90 and recall of 0.88, indicating its
reliability in diagnosing Alzheimer’s cases.
• Provide predictions within seconds, making it suitable for real-
time applications.
The web application effectively enables users to interact with the
system, ensuring ease of use for non-technical individuals.
Future Scope
The project lays the foundation for advanced Alzheimer’s diagnostic
tools but has room for further improvement:
1. Expanding the Dataset: Including more diverse brain scans
from various demographics to improve the model's
generalizability.
2. Multimodal Inputs: Incorporating other diagnostic data such as
genetic information and cognitive test scores.
3. Cloud Deployment: Hosting the application on cloud platforms
to enhance accessibility and scalability.
4. Real-Time Feedback: Integrating feedback mechanisms for
continuous learning and improvement.
5. Mobile Application: Developing a mobile-friendly version to
make the tool accessible to a broader audience.
Conclusion
This project demonstrates the potential of machine learning in
revolutionizing healthcare diagnostics. By leveraging deep learning
techniques and a user-friendly web interface, the system provides a
reliable and accessible method for early detection of Alzheimer’s
disease. The integration of Flask for backend and frontend
development ensured seamless interaction, making the application
practical for real-world use. With further enhancements, this tool can
play a pivotal role in combating Alzheimer’s by facilitating timely
diagnosis and intervention.