Customized DL Model for Alzheimers Detection from MRI Images

This Project Aims to Develop a Customized Deep Learning Model which Detects 4 Classes of Alzheimer (Non Demented, Very Mild Demented, Mild Demented and Moderate Demented) from MRI Images. The goal is to improve early detection and diagnosis of Alzheimer’s disease using non-invasive and accessible imaging technology

Models and Files from this project

- Model Model_1
- Model Model_1.h5
- Model Model_2
- Model Model_2.h5
- Model Model_Inception
- Model Model_Inception.h5
- Variables.npz

Data

Goal is to Train a Deep Learning Model on this type of Data. The trained model will be able to detect different classes of alzheimer from MRI Images.

Model Model_1

- Only Custom Callbacks
- Dropout after conv_block(64)

Results from Model_1

Model Model_2

- Custom + Reduce LR on Plateau Callbacks
- No Droput After conv block(64)

Results from Model_2

Model Inception V3

Results from Inception V3

Built With

• classification
• data-analysis
• data-science
• deep-learning
• image-processing
• jupyter-notebook
• machine-learning
• python
• smote
• streamlit