- Steve Piepers Step library for WebGPU
- https://github.com/pieper/step)
- https://github.com/OHIF/ohif-step
- https://github.com/OHIF/Viewers/pull/1237/files
- https://github.com/JamesAPetts/Viewers/blob/8b573c4c84326296c2b015be51f06bdafb3b0c4c/extensions/cornerstone-spherical-growcut/src/tools/segmentation/SphericalGrowCutSegmentationTool.js
- https://github.com/JamesAPetts/Viewers/blob/8b573c4c84326296c2b015be51f06bdafb3b0c4c/extensions/cornerstone-spherical-growcut/src/tools/segmentation/strategies/growCutFromCircle.js
- De-identification for radiology/dicom and pathology (rahul)
- Working on NHI project to de-identify DICOM and digital pathology
- Focus is on burned in PHI
- Tools to automatically detect and de-identify metadata and pixels
- General thoughts
- Huge variation in what needs to be identified based on the region, research project, etc
- Opt in instead of opt out
- Assume that you will miss something
- Be able to re-process the data when issues are found
- Be able to keep track of each change in de-identification scheme
- Be able to keep track of who accessed what datasets
- NIH has requriements about removing all PII from an x-ray
- Pacemaker serial number?
- Google OCR (Not the de-id in the healthcare api)
- works pretty well
- Facial recognition
- https://frederick.cancer.gov/collaborations/business-opportunities/aimi-datasets
- Patient controlled/managed data
- Joël Spaltenstein https://www.agoracare.ch/en/
- Adrian Gropper: https://hieofone.com/
- https://www.linkedin.com/posts/bartdewitte_google-sued-for-using-the-nhs-data-of-16-activity-6933678532317360128-eM0t/
- Portugese company that will link hospitals to allow patient sharing
- Joao update on his PHD/AI Training with MR