Waymo is proud to host and participate in multiple sessions at the 49th Conference on Computer Vision and Pattern Recognition (CVPR) next week
Waymo is proud to host and participate in multiple sessions at the 49th Conference on Computer Vision and Pattern Recognition (CVPR) next week. We will present some of our recent state-of-the-art work in autonomous driving research—one of the most complex and practical applications of computer vision.
Last year alone, Waymo-affiliated researchers published 14 papers—and this year’s number is set to be even higher! As we gear up for CVPR, we are also excited to launch a brand new Waymo Research Website, in addition to our Github and the Waymo Open Dataset sites, as a destination to make our state-of-the-art breakthroughs more easily discoverable for other researchers.
We strive to push the boundaries of cutting-edge research, as we pursue our mission to make it safe and easy for people and things to get where they’re going. As we tackle some of the hardest problems in machine learning and see the potential our discoveries and inventions have for real world applications, we are excited to share insights with the research community.
Waymo at CVPR 2022
If you are participating in CVPR this year, either in person or virtually, please join us for the following sessions and stop by the Waymo booth to meet our team.
- On June 19, at the LatinX in CV (LXCV) Research workshop, Xinchen Yan will hold a tutorial session on synthetic camera data generation for autonomous driving.
- On June 20, at the Workshop on Autonomous Driving, the Waymo Research team will share the results from this year’s Waymo Open Dataset Challenges and unveil some exciting new additions to this industry leading dataset.
- On June 22, a joint team from Waymo and Google Research, including Vincent Casser, Xinchen Yan, Sabeek Pradhan, and Henrik Kretzschmar, will present BlockNeRF, a novel method for large-scale scene reconstruction based on camera images. Earlier this year, the team used this technique to recreate an entire neighborhood of San Francisco from 2.8 million images—the largest such NeRF-based 3D reconstruction, to date. Waymo also recently released the Waymo Block-NeRF Dataset, one of the datasets used in the experimental evaluation presented in the paper, so other researchers can apply their scene reconstruction methods to the dataset, too.
- On June 24, Waymo research team members will present their work in a poster session on a novel data-driven range image compression algorithm, nicknamed RIDDLE (Range Image Deep Delta Encoding). As lidar become more powerful with increasing resolutions, there is a need for lidar data compression which can lower the costs of data storage and transmission. The proposed method demonstrates a large improvement in the compression quality compared to previous algorithms.
Waymo Open Dataset 2022
The Waymo Open Dataset is one of the ongoing contributions to the research community we are most proud of at Waymo. This year’s Challenges have been a great success with record-high participation with over 1,700 valid submissions, including entries from more than 20 universities worldwide. Our team was delighted to see the variety of creative approaches the research community contributed to this year’s perception and prediction challenges, and the high quality of the submissions. Many of the submissions are pushing the state-of-the-art in the respective research fields.
- Review the recent expansions of Waymo Open Dataset features
- Introduce 2D Video Panoptic Segmentation Labels, which adds per-pixel semantic and instance segmentation labels to the existing camera images for 28 classes
- Announce the winners and discuss the winning methods from the four 2022 Waymo Open Dataset Challenges: 3D Camera-Only Detection, 3D Semantic Segmentation, Motion Prediction, and Occupancy and Flow Prediction
- We have also released a new paper, LET-3D-AP, describing the metric used to evaluate submissions to this year’s Waymo Open Dataset Camera-Only 3D Detection Challenge