A Neurosurgical Instrument Segmentation Approach to Assess Microsurgical Movements.

The ability to recognize anatomical landmarks, microsurgical instruments, and complex scenes and events in a surgical wound using computer vision presents new opportunities for studying microsurgery effectiveness. In this study, we aimed to develop an artificial intelligence-based solution for detecting, segmenting, and tracking microinstruments using a neurosurgical microscope. We have developed a technique to process videos from microscope camera, which involves creating a segmentation mask for the instrument and subsequently tracking it.