Detection2 Benchmark

VISCERALlesionilluvert

The Detection2 Benchmark is a continuously running Benchmark for Lesion Detection in 3D medical images.

Data

We distribute medical imaging data (CT, MRI, w & w/o contrast enhancement) that contains various lesions in anatomical regions such as the bones, liver, brain, lung, or lymph nodes. There are overall 1609 annotated lesions in the dataset.

Task

You get imaging data and annotations in the form of lesion center position, and for large lesions, annotations that indicate the radius. You should train your lesion detection algorithms on this data.

You should then run your trained detection algorithms against test data (also available) and evaluate the results using the provided evaluation tool (EvaluateSegmentation). This will evaluate the detection performance against expert annotations of lesions.

Documents and Resources

Register for a benchmark account at the VISCERAL registration website.  Choose "Detection2" as the Benchmark for which to register (Virtual Machines will not be used in this Benchmark, so the VM selection can be left as is). As a next step, a participation agreement must be signed and uploaded. Once this is done and accepted by the organisers, access will be granted to the participant dashboard.

The participants have access through the participant dashboard to instructions for downloading the data via ftp and a data handling tutorial.  

The Detection2 Benchmark Specification can be downloaded (currently v1.0 of 20150619)

The latest version of the EvaluateSegmentation tool can always be downloaded from: https://github.com/codalab/EvaluateSegmentation (the version including the detection evaluation metrics will be available in the first week of July 2015).

Questions?

If you have any questions, please feel free to get in touch: hanbury@ifs.tuwien.ac.at