Project InnerEye is a Microsoft Research initiative that uses AI to analyze 3D medical scans — primarily CT scans used for cancer radiotherapy planning — and automatically highlight the areas that need treatment. Peer-reviewed research found it can complete this task up to 13 times faster than a clinician doing it manually, while staying within the range of human expert accuracy.

Developed over more than a decade by Microsoft Research's Cambridge (UK) lab, Project InnerEye uses deep learning models — a type of machine learning inspired by how the brain processes information — to perform automatic segmentation of medical images, meaning it identifies and outlines specific organs and tumors in 3D scans. The open-source InnerEye Deep Learning Toolkit was released under an MIT license in September 2020, built on PyTorch and Azure Machine Learning. A key real-world deployment is OSAIRIS at Cambridge University Hospitals NHS (UK), which uses InnerEye-based software to reduce wait times for cancer radiotherapy patients. Results were published in JAMA Network Open in 2020. The GitHub repository for the toolkit showed reduced activity after 2022, and the current status of the project under Microsoft Health Futures should be confirmed.