Brain MRI classification

Data challenges, in the form of an open competition, enable a community of data scientists to compete on single problems. A collective of researchers working on a challenge typically finds a better solution than one scientific group or company by itself, in what is essentially crowd-sourcing. Open challenges have resulted in several scientific breakthroughs and improve the state-of-the art in medical image processing. In a challenge, many algorithms are tested on a dataset. This makes algorithms easier to compare and transparent. The benefits of challenges extend beyond the solution or winner, as challenges foster data sharing, collaboration and transparent science. However, it is currently not easy to host an open challenge with sensitive data from the medical domain.

In the context of the “Nationale Wetenschapsagenda” (NWA) roadmap for Value Creation through Responsible Access to and use of Big Data (VWdata) a scientific challenge is being created. The scientific challenge contains, among others, a medical challenge. The medical challenge is to classify clinically acquired brain MRI scans as normal or abnormal. This is a highly important clinical task in neuroradiology. At Radboudumc alone, over 2.000 exams are performed annually to resolve this question for patients with a wide variety of non-specific complaints.

The RTC Deep Learning is collecting data and setting up a challenge for the classification of 3D brain MRI scans.

People

Eric Postma

Eric Postma

Professor

Artificial Intelligence, Tilburg University

Silvan Quax

Silvan Quax

['Research Scientist', 'RTC Deep Learning']

Ajay Patel

Ajay Patel

Coordinator RTC Deep Learning

Bram van Ginneken

Bram van Ginneken

Professor

Diagnostic Image Analysis Group

James Meakin

James Meakin

Lead Research Software Engineer

DIAG Research Software Engineering