Motional, the new joint venture from Hyundai Motor Group and Aptiv, has announced an expansion to nuScenes. The public dataset, which debuted last year, teaches autonomous vehicles how to navigate changing road environments. With the recent expansion, nuScenes includes 1.4 billion annotated LiDAR points and nearly 100,000 annotated 2D images.
The public dataset became available for the first time in March of 2019. It consisted of 1,000 different street and urban scenes throughout Boston and Singapore. The scenes, comprised of millions of photos and data-points collected from full sensor suites, were then meticulously hand-annotated.
Platform of Collaboration
Although the project was just starting at the time, the idea was to create a culture of data sharing and collaboration. “Safety transcends competition,” said Karl Iagnemma, Motional President and CEO. “The belief that passenger safety must take priority over any competitive advantage is at the heart of nuScenes.”
Join the ADAS Insiders
Join other industry experts in the AutoVision News ADAS Insiders Community today for early and exclusive access to industry news, events, business directory, and much more.
According to the company, nuScenes has been cited in more than 250 scientific papers and referenced by over 8,000 researchers since its release in March of 2019. “We’ve been delighted to see so many peers follow suit and release their own datasets, all for the betterment of the industry,” Iagnemma said.
The recent expansion includes two important elements: nuScenes-lidarseg and nuImages.
nuScenes-lidarseg applies lidar segmentation to the original Singapore and Boston driving scenes. The LiDAR segmentation, as described by Motional, provides a more accurate picture of a vehicle’s surroundings than the original nuScenes’ bounding boxes. Motional says the inclusion of nuScenes-lidarseg, with its 1.4 billion annotated LiDAR points, will help researchers quantify challenges like LiDAR point cloud segmentation and foreground extraction.
The second expansion, nuImages, is a collection of 100,000 annotated images representing a wide range of unpredictable and challenging driving conditions. nuImages was created in response to user demand to address difficult road challenges like large intersections and changing weather.
How to Access nuScenes?
nuScenes is free for academic use, and licensing is available for commercial purposes. Scale AI provided the annotations. Along with launching nuScenes last year, Motional also contributed to the Safety First for Automated Driving whitepaper.