Large scale publicly accessible image archives facilitate reproducible research and deep learning. Data quality and provenance tracking are primary concerns. Posda is a set of open-source scalable curation tools that we developed to address these concerns. Posda is the primary curation tool used by TCIA to ensure submitted DICOM objects are free from PHI, are valid DICOM objects, and are scientifically useful.

Learning Objective: Describe the steps necessary to fully de-identify DICOM image objects for public release

Authors:

William Bennett (Presenter)
University of Arkansas for Medical Sciences

Kirk Smith, University of Arkansas for Medical Sciences
Quasar Jarosz, University of Arkansas for Medical Sciences
Sonya Utecht, University of Arkansas for Medical Sciences
Fred Prior, University of Arkansas for Medical Sciences

Presentation Materials:

• Abstract/Manuscript - 1