pyfroc 0.1.0

python toolkit for FROC analysis

pyfroc

Python framework for FROC analysis

About

What pyfroc does

• Improve FROC analysis procedure.
• Manage responses of raters.
• The responses can be made using segmentation function of 3D Slicer.
• Evaluate responses and devide them into true positive or false positive automatically.
• Build a xlsx file for the RJafroc, a R library which runs statistical tests of AFROC (alternative Free-response receiver operating characteristic) analysis.
• Write images of responses with paired lesions (if exists).

What pyfroc doesn't

• Statistical analysis of JAFROC. This is out of scope of pyfroc. Use RJafroc for statistical analysis.
• FROC analysis including multi-modality references because pyfroc doesn't implement an algorithm to match intermodality lesions.

Table of contents

Use case

pyfroc is designed for specific scenarios of FROC analysis. pyfroc supports only one modality for reference lesions.

Example scenario #1

• Compare diagnostic performance between radiologists and AI.
• Using a specific series to record responses.

Example scenario #2

• Compare a standard MRI protocol with an abbreviated protocol.
• Using same series to record responses.

Example scenario #3

• Compare images reconstructed using an advanced method with images reconstructed using conventional method.
• Using either series to record responses.

Instalation

pip install pyfroc

Tutorial

Use pyfroc with 3D Slicer

License

GPLv3