Computer Vision News - December 2021

18 AI Spotlight News Computer Vision News has found great new stories, written somewhere else by somebody else. We share them with you, adding a short comment. Enjoy! ‘Cooperative Perception’ Could Help Autonomous Vehicles See Pedestrians Around Corners New technology developed by researchers in Australia could help autonomous vehicles steer clear of pedestrians and cyclists without a direct line of sight. Three researchers from the Australian Centre for Field Robotics at the University of Sydney claim they can “substantially improve the efficiency and safety of road transportation.” One of the “intelligent roadside units” was even able to “see” a “visually occluded pedestrian behind a building” seconds before it would have otherwise. The image here shows a cooperative-perception enabled vehicle detecting a cyclist behind a bus. Read More A New Camera Can See Through Almost Anything, Including Human Tissue and Bones Adifferent camera: this toohas theability to “see” through things. Developed by Northwestern Engineering researchers, this new HR camera can see around corners and throughhumanskinandeven bones. It also has the potential to image fast-moving objects such as speeding cars or even the beating heart. This non- line-of-sight (NLoS) imaging camera is so precise that it even captures the tiniest capillaries at work. It intercepts the scattered light coming from an object to reconstruct the inherent information about its time of travel and reveal the original object. Read More Facebook battles the challenges of tactile sensing Facebook (or Meta if you want) wishes to lead the rush for touch sensing andhas developeda full set of hardware and software that provide this function, an emerging need in the robotics field: it aims to understand and replicate human-level touch in the physical world, providing information about temperature, texture, weight and more. Together with the Carnegie Mellon University, they presented a synthetic “skin” called Reskin, which will be an inexpensive and replaceable solution for long-term use, employing an unsupervised learning algorithm to help auto-calibrate the sensor. Read More