Computer Vision News - March 2022

43 Percutaneous Nephrolithotomy (PCNL) and lithotripsy and suction tubes to view potential pathways and directions.  Image Registration: A combination of advanced Artificial Intelligence and classic computer-vision algorithms are trained to accurately register the fluoroscopy and ultrasound images with the pre-op scan, overlaying the detailed planning map on the real-time images. This tool provides needle position and direction within the 3D information of the pre-op scan, adding a depth dimension in real-time. This method may obviate the use of fluoroscopy, as it maximizes the use of ultrasound, thus reducing radiation exposure during the treatment. Needle Tracking: State-of-the-art tracking modules follow the needle and provide an accurate estimation of the needle position within the images. This provides continuous assessment of the needle position and allows correction if needed. Combining these tools provides the surgeon with real-time, precise information about the stone location and the needle position throughout the procedure , reducing risk of complications and speeding up the surgery. This set of tools can be introduced into robotic systems , thus improving PCNL outcomes even more by using AI and the robot’s precision.  RSIP Vision has a proven track-record of implementation of segmentation, registration, and tracking algorithms. Our experienced clinicians and engineers collaborate to provide a quick and accurate solution toany clinical need, usingadvanced AI technology. Learn more about their AI solutions for urology. and the operating-room staff as it increases the risk for future disease, so a minimal radiation dosage is required. To address these challenges, RSIP Vision presents a set of AI tools : Procedure Planning: The pre-op scan contains both high resolution and 3D information. RSIP Vision ’s dedicated deep-learning algorithms can detect and segment the kidney stones , as well as surrounding anatomic structures. Using this output, a detailed surgical plan can be developed - from point of incision, to needle angle and depth penetration, and even further planning of specific methods for stone removal. The latter may include insertion of 3D models of standard needles