Quality assessment algorithm evaluation

This repository currently contains "Error versus Discard Characteristic" (EDC) Python example/utility code, which can be used to evaluate quality assessment algorithms independent of the biometric modality.

• edc.py: Contains general code to compute EDC curve data, including "partial Area Under Curve" (pAUC) values. The only required Python package for this module is numpy.
• example.py: Contains a small CLI example that produces an EDC plot using the plotly package. The plot will be opened in the default browser. This example also includes plot utility functions.
• example_data.json: Default example input data used by example.py.

To quickly create a new Anaconda Python environment for example.py, run:

conda create --name edcexample python=3.9
conda activate edcexample
pip install -r requirements.txt
python example.py

There are various CLI options in the example, see python example.py --help.

A related "Face Image Quality Assessment Toolkit (fiqat)" with various additional features can be found here: https://share.nbl.nislab.no/g03-03-sample-quality/face-image-quality-toolkit