Projects » VISCERAL
VISCERAL is a support action that will organize two competitions on information extraction and retrieval involving medical image data and associated text that will benchmark the state of the art and define the next big challenges in large scale data processing in medical image analysis.
The increasing amounts of medical imaging data acquired in clinical practice hold a tremendous body of diagnostically relevant information. Only a small portion of these data are accessible during clinical routine or research due to the complexity, richness, high dimensionality and size of the data.
There is consensus in the community that leaps in this regard are hampered by the lack of large bodies of data shared across research groups and an associated definition of joint challenges on which development should focus. VISCERAL will provide the means to jump-start this process with two competitions (1) providing access to unprecedented amounts of real world imaging data annotated through experts and (2) using a community effort to generate a large corpus of automatically generated standard annotations. The goal of the project is to formulate relevant and challenging tasks, to provide the necessary data for research and evaluation, and to conduct competitions for identifying successful computational strategies and highlighting directions of future research.
To this end, VISCERAL will conduct two competitions. The first competition will focus on automatic identification, localization and segmentation of anatomical structures in medical imaging data, the second competition will comprise retrieval tasks that aim at identifying similar cases relevant for diagnosis. In addition to the direct evaluation, the project will result in two data corpora - one gold corpus of expert manual annotations, and a silver corpus that will be computed from the competition entries. Both data sets will be made available to challenge participants and, afterwards, to the scientific community.