Waymo LLC, Alphabet Inc.’s self-driving car subsidiary, hopes to advance the state of artificial intelligence by providing researchers with new data for their projects.

The group today released a massive corpus of sensory information collected by its autonomous vehicles. The repository includes 1,000 video segments recorded under a variety of driving conditions in both day and night.

Waymo’s dataset doesn’t seem very large at first glance: The 1,000 segments are all 20 seconds long, which adds up to about five and a half hours of footage. But every segment consists of 200,000 individual frames that each constitutes a useful data point in its own right. Waymo furthermore individually labelled the vehicles, pedestrians, cyclists and signage that appear in the video, overlaying no fewer than 13.2 million data tags on top of the frames.

“We believe it is one of the largest, richest, and most diverse self-driving datasets ever released for research,” Waymo Principal Scientist Drago Angelo wrote in a blog post.

Another factor that makes the data valuable for researchers is the method with which it was assembled. According to Waymo, the driving segments stitch together footage from five different cameras on the front and sides of its self-driving cars, providing a 360-degree view. The files also include an added layer of environmental data gathered by the five lidar sensors in each vehicle.

“When it comes to research in machine learning, having access to data can turn an idea into a real innovation,” Angelo wrote. “This data has the potential to help researchers make advances in 2D and 3D perception, and progress on areas such as domain adaptation, scene understanding and behavior prediction.”

These are fields with practical applications that extend far beyond self-driving cars. Waymo’s dataset can potentially lend itself just as well to developing computer vision algorithms for robots, drones and video analysis software to name a few applications. The only restriction is that the licensing terms limit use to non-commercial projects.

Waymo’s dataset is the latest addition to the burgeoning ecosystem of autonomous driving tools that has formed within the broader open-source community. Uber Technologies Inc., Scale Labs Inc.  and a number of players in the self-driving car race have released research datasets too, along with related tools such as visualization software. 

Photo: Waymo