Computer Vision News - January 2024

Computer Vision News 38 NeurIPS Paper Nina Montaña Brown is a finalyear PhD student at UCL jointly at the Centre for Medical Image Computing (CMIC) and the Welcome/EPSRC Centre for Interventional and Surgical Sciences (WEISS). She speaks to us about her NeurIPS paper proposing the first large-scale dataset of 3D rendering assets designed to simulate computer vision data for synthetic tasks about surgery. The ever-increasing need for data has driven advancements in predictive models to perform specific tasks, particularly within the realm of computer vision. We have seen incredible technological advances in self-driving cars, for example, which can learn to navigate in an open-ended environment based on imaging and depth information. “Increasingly, we’re seeing algorithms being trained on synthetic data,” Nina begins. “It’s quite similar to when you play a video game. You can simulate a camera and its position in a synthetic environment that’s made up of 3D objects and their textures and the lighting conditions, and that gives us an output of an image of that environment.” Matias Valdenegro SARAMIS: Simulation Assets for Robotic Assisted and Minimally Invasive Surgery