A research team led by the Mohamed bin Zayed University of Artificial Intelligence has developed Janus AI, a video-to-video retrieval system designed to help surgeons find similar procedures while an operation is taking place. The system won first prize at the Robotic Urology Breakthrough Innovation Challenge at the European Robotic Urology Symposium.
Janus AI compares video captured during surgery with a library of stored procedures. The system represents short video segments through features extracted by a neural network and then searches for similar clips. Because the library features are precomputed, the live operating-room video is the main element that needs to be processed at query time. MBZUAI says the similarity search can typically be completed in milliseconds.
The system is intended as a reference and educational tool. Surgeons can retrieve a comparable 30-second segment and continue into the longer procedure to see how the operation unfolded. For trainees and less experienced surgeons, the ability to review carefully selected examples may help build understanding of the stages of a robotic operation.
The research sits at the intersection of computer vision, surgical robotics and clinical education. Its practical value will depend on how well it performs across different procedures, camera angles, institutions and patient conditions. A reference library also requires careful curation: videos need appropriate consent, consistent annotation and strong privacy protection. The system should assist clinical judgement rather than replace the surgeon’s responsibility for the patient.
The project is a collaboration between MBZUAI, the University of Illinois Chicago and the University of Turin. The team plans to improve the architecture, expand the library with additional surgical videos, involve graduate students in development and validation, and continue work toward clinical translation.
Janus AI is not presented as a deployed autonomous surgical system. It is a retrieval and decision-support technology under further development. Its significance for the UAE AI ecosystem is that it shows how local research institutions are applying advanced vision models to high-consequence professional workflows, where speed, relevance, validation and human oversight all matter.
