A Bio Image Analysis workflows benchmarking platform.

Getting started is easy!

BIAFLOWS is a web platform centralizing Bio Image Analysis (BIA) workflows to automatically extract biological objects annotations from microscopy images. The workflows can be launched online on a predefined set of annotated images illustrating classical BIA problems (e.g. Nuclei segmentation), and the results visualized or quantitatively compared from documented benchmark metrics. To ensure full reproducibility, the workflows are packaged with their execution environments as versioned Docker images compiled from open source code repositories.

Read the article, or browse the following sections to learn more about BIAFLOWS.


How to benchmark BIA workflows and why is it critical?

Video tutorial

Demo on how to use BIAFLOWS online instance


Access BIAFLOWS online instance


Access BIAFLOWS sandbox

User Documentation

A User guide to BIAFLOWS

Community forum

Where to ask questions and get answers on anything related to BIAFLOWS

Installing and populating BIAFLOWS locally

Install a local BIAFLOWS server to process your own images with public workflows

Creating a BIA workflow and adding it to a BIAFLOWS instance

Package your own workflow to make it available from BIAFLOWS

Creating a BIA workflow and adding it to a BIAFLOWS instance (Developer guide)

Provide extra information and the procedure to build and debug workflow Docker images locally

Problem classes, ground truth annotations and reported metrics

The different BIA problem classes, computed metrics and annotations encoding

Executing a BIAFLOWS workflow without BIAFLOWS server

Process a folder of local images with a BIAFLOWS workflow

Access BIAFLOWS from a Jupyter notebook

Access BIAFLOWS from a Jupyter notebook to download images and process them (launch Binder from linked GitHub repository)