Retinal blood vessels segmentation

Detection of retinal blood vessels using classical methods (edge detection using filters) and deep neural network methods (UNet).

Web application

Windows-64

1. Download current UNet release from Releases.
2. Place downloaded .pth file in models folder.
3. Execute commands:

conda create --name <env> --file .\requirements.txt
conda activate <env>
python main.py

Web application allows the user to upload their own images and see the results.

Jupyter Notebook

All core functionalities of the project are located in the notebooks folder.

Built with

• Python 3.9
• PyTorch
• OpenCV
• Flask

Contact

[email protected]