Sign up for the NCI Medical Image De-Identification Benchmark (MIDI-B) Challenge if you’re a developer (independent, academic, or industrial) who wants to assess your software tool’s performance with rule-based de-identification of DICOM images (DICOM being “Digital Imaging and Communications in Medicine”). Automated image de-identification methods that preserve the research utility of the data are of particular interest for this challenge.

How the Challenge Will Work

The MIDI-B Challenge will use subsets of imaging data from the greater MIDI data set. The MIDI data set is a large and diverse multi-modality set of medical images selected from public collections in NCI’s The Cancer Imaging Archive (also available in the Imaging Data Commons) and infused with synthetic identifiers.

After completion of this challenge, you’ll receive an invitation to contribute as a co-author to the MIDI-B Challenge report. The challenge organizing team will lead the development of the report manuscript and identify a peer-reviewed journal to which to submit the publication.

The three top-ranked teams will receive an invitation to present their algorithms and results at the Medical Image Computing and Computer Assisted Intervention (MICCAI) MIDI-B Challenge virtual session on October 10, 2024. MICCAI Satellite Events require registration.

Important 2024 Dates

• May 10–Aug. 15: Registration open
• May 10–Aug. 3: Training phase
• Aug. 4–30: Validation phase
• Sept. 4–6: Test phase
• Oct. 10: MIDI-B virtual session at MICCAI 2024

Additional Information

Data sharing promotes the development of analytic tools, including artificial intelligence. This supports cancer research. A fundamental requirement for sharing medical images through public repositories is removing patient identifiers while following best practices to preserve the research value of the data.  

The MIDI-B Challenge is one of the Satellite Events of the 27th International Conference on MICCAI 2024 in Marrakesh, Morocco. NCI CBIIT leads the organization of MIDI-B in collaboration with the National Heart, Lung, and Blood Institute, Sage Bionetworks, Frederick National Laboratory, University of Arkansas for Medical Sciences, Ellumen Inc., PixelMed Publishing, and Deloitte.