Waymo Open Dataset Challenges put autonomous vehicle engineers to test

Waymo today extended the suspension of its autonomous vehicle programs to now include fully autonomous vehicles. Earlier in the week, Alphabet-owned Waymo paused operations of autonomous vehicles that had safety drivers on board. This is due, of course, to the ever-growing COVID-19 pandemic and the novel coronavirus.

Like many companies, Waymo is looking to keep people quarantined at home busy and get their minds off of COVID-19. And Waymo is even offering up cold, hard cash with the Waymo Open Dataset Challenges.

“Even as COVID-19 continues to develop, we are committed to fostering an environment of innovation and learning – one that can continue to grow and thrive in our temporarily virtual world,” wrote Drago Anguelov, Principal Scientist, Waymo.

Researchers can partake in these five competitions based on the data gathered by Waymo’s driverless fleet.

• 2D Detection: Given a set of camera images, produce a set of 2D boxes for the objects in the scene.
• 2D Tracking: Given a temporal sequence of camera images, produce a set of 2D boxes and the correspondences between boxes across frames.
• 3D Detection: Given one or more lidar range images and the associated camera images, produce a set of 3D upright boxes for the objects in the scene.
• 3D Tracking: Given a temporal sequence of lidar and camera data, produce a set of 3D upright boxes and the correspondences between boxes across frames.
• Domain Adaptation: Similar to the 3D Detection challenge, but we provide additional segments from rainy Kirkland, Washington, 100 of which have 3D box labels.

The Waymo Open Dataset Challenges are now open and run through May 31. Prizes will be given to the top submissions in each of the five challenges. First place will walk away with $15,000, second place will receive $5,000, and third place will take home $2,000.

Winners of the Waymo Open Dataset Challenges will have the chance to showcase their work at CVPR 2020 in Seattle on June 14. However, the computer vision event may be in jeopardy due to ongoing concerns surrounding COVID-19, so Waymo said other venues will be in play if necessary.

Waymo Open Dataset

The Waymo Open Dataset, which is available for free, is comprised of sensor data collected by Waymo self-driving cars. It covers a variety of environments, from dense urban centers to suburban landscapes, and includes data collected during day and night, at dawn and dusk, in sunshine and rain. Waymo updated the dataset this week. Here’s a breakdown of the Waymo Open Dataset:

Size and coverage: Data from 1,950 driving segments. Each segment captures 20 seconds of continuous driving, corresponding to 200,000 frames at 10 Hz per sensor. Such continuous footage gives researchers the opportunity to develop models to track and predict the behavior of other road users.

High-resolution, 360° view: Each segment contains sensor data from five high-resolution Waymo lidars and five front-and-side-facing cameras.

Labeled Data: The dataset includes labels for four object classes: vehicles, pedestrians, cyclists, and signs. There are 12.6 million 3D bounding box labels with tracking IDs on LiDAR data and 11.8 million 2D bounding box labels with tracking IDs on camera data.