
GRIFD LSTM GUI Version - README
---------------------------------
DESCRIPTION:
This version includes a GUI-based folder selection and implements an LSTM-based deep learning model for DR detection.

REQUIREMENTS:
- MATLAB R2021 or newer
- Deep Learning Toolbox
- Image Processing Toolbox

INSTRUCTIONS:
1. Download and extract the EyePACS or Mendeley fundus dataset.
2. Place all `.jpg` images in one folder (can include normal and DR images).
3. Run the script `grifd_lstm_gui.m` in MATLAB.
4. A folder selection dialog will appear – select the folder containing fundus images.
5. The model will preprocess, extract features, train the LSTM, and show training progress.
6. Dummy labels are used (even=normal, odd=DR) – replace with true labels if available.

NOTES:
✓ Implements sequential modeling using LSTM (RNN)
✓ Uses Local Binary Patterns (LBP) and edge features
✓ GUI interface for dataset folder selection
✓ Compatible with EyePACS, Mendeley, and similar datasets

For research-grade use, replace dummy labels with annotated DR stage labels.
