Updated: 21 December, 2018
RECOMIA's research portal is open to academic researcher. We continously improve the tools available together with our research partners. Your data will remain private for the duration of your research project and you can easily control who and when people should get access to your data. When your research is concluded we appreciate the opportunity to use your data in a fully de-identified form to train new deep learning models, which will help us to improve our research forum - for the benefit of all academic researchers.
PSEUDO DE-IDENTIFIED DATA
Export your imaging data from your institutions' PACS system.
All data is automatically de-identified locally at the submitting site before data is uploaded via secure communication to RECOMIA's research forum. Hence, no sensitive data will leave your hospital's closed network. Our de-identification process follows industry-best practice. Learn more
Pseudo de-identified data
The de-identification scripts exclude ID mapping between the original Patient ID and the ID that the images will have within RECOMIA. Any manual mapping that is performed locally at the submitting site, is the sole responsibility of the researcher, and RECOMIA never sees the original Patient ID. Hence, data is pseudo de-identified, as the data can no longer be attributed to a specific data subject without the use of additional information (only avilable at the submitting site).
The manual labelling process of medical images is tedious, time consuming and complex, thus accurate dataset annotation can be a major barrier to new scientific discoveries. Our research forum has build-in convolutional neural networks that enables academics to annotate their imaging data (CT, SPECT/CT, PET/CT, MRI, etc.) with many unique ROI automatically, and get instant quantification results. Learn more
Quantification results are avilable for download, and measures of quantification is added continously in close collaboration with our research partners.
Fully de-identified data
After data processing, the ID used to identify the images on the research forum are transformed, rendering the images fully de-identified in such a way that the data subjects are not or no longer identifiable. In this process, we strip all images of any identifiable information, making it impossible to derive insights on a discreet individual, even by the party that is responsible for the anonymization or by the sumitting party.
Training deep learning models
Fully de-identified data is used to train new deep learning models.
New and improved tools openly avilable for academic researchers on the research forum.