Reported 8 months ago
During the AI City Challenge at the International Conference on Computer Vision and Pattern Recognition (CVPR) on June 19, 2024, NVIDIA presented its largest indoor synthetic dataset to aid researchers and developers in advancing solutions for smart cities and industrial automation. This dataset allows researchers and developers to test models and quickly create applications and workflows based on universal scene descriptions (Open USD) through the NVIDIA Omniverse platform. By utilizing digital twins to simulate environments, tasks such as optimizing factory operations or warehouse logistics can be efficiently arranged, improving efficiency and ensuring the smooth operation of robots and autonomous vehicles. The synthetic dataset, generated using NVIDIA Isaac Sim platform's Omniverse Replicator tool, offers a range of virtual environments and scenarios for training and simulating interactions between robots and their surroundings. Participating institutions include National Yang Ming Chiao Tung University, Australian National University, UAE Mobile Research Center, IIT Kanpur, Iowa State University, Johns Hopkins University, University of Santa Clara, UAE University, SUNY Albany, and Toyota, among others.
Source: YAHOO