Overview

Prerequisites

This project uses these:

• Python 3 as programming language
• Docker for environment management and reproducibility
• DVC for data and model versioning
• uv/pyproject for environment management
• MLflow as tracker
• WandB for experiment tracking and visualization (TBD)
• PyTorch as DL framework
• Makefile as command line interface
• FastAPI (TBD)
• Streamlit (TBD)
• pytest and ruff for testing and linting (TBD)

Datasets Used

ROI extraction

• UBIPr dataset of periocular images: https://socia-lab.di.ubi.pt/~ubipr/

Blood vessel extraction

Experiment tracking

Data/model versioning