RedBrick AI

RedBrick AI is a browser based annotation platform for medical imaging, used to build the ground truth datasets that radiology AI models are trained and validated on. Teams upload DICOM studies, label them with segmentation and classification tools, and run the results through review and consensus steps before exporting the data for model development. It is built for AI teams training radiology models, medical imaging companies that need annotated datasets, and hospitals and research groups running clinical validation studies, including those that must document their workflow for FDA or CE submissions. The annotation toolkit includes an automated segmentation tool built on Meta's Segment Anything model, region growing, thresholding, Boolean operations, three dimensional contouring and interpolation between slices. The viewer renders volumes in 3D with maximum intensity projection, oblique planes and photorealistic rendering. Project management covers multi step review, consensus between annotators and quality control, and the platform scales from a couple of annotators to hundreds. Data management adds search and filtering over DICOM tags, custom validation checks and audit trails, and an API and SDK give programmatic control with the option to keep images in the customer's own storage. Pricing is not published. Prospective customers request access through the website and receive a quote. Nothing is installed locally: annotation, 3D viewing and review all run in the browser, which is what allows a distributed group of radiologists to work on the same project.