We are excited to announce the launch of the third Robotic Vision Scene Understanding Challenge (RVSU), being run at the 2022 edition of the CVPR Embodied AI workshop.
The RVSU challenge, which has $2500USD prize money on offer, prompts researchers to develop robotic vision systems that understand both the semantic and geometric aspects of the surrounding environment. The challenge consists of 6 tasks, featuring multiple difficulty levels for object-based semantic simultaneous localisation and mapping (SLAM) and scene change detection (SCD).
In order for robots to operate in household environments, they need to understand what objects are in their environment and where they are. This is the first challenge captured in our semantic SLAM tasks, where a robot must explore an environment, find all objects of interest, and add them to a 3D map. The SCD task takes this a step further, asking a robot to report changes to the objects in an environment at a different point in time. See the video below for a full overview of the RVSU challenge:
Our established BenchBot framework powers the RVSU challenge. BenchBot gives participants access to a rich ecosystem of photorealistic 3D simulations, ray-traced lighting and reflections, physics engine enabled robots, simulated sensorimotor data, and repeatable experiments. We have released a new version of BenchBot, which use NVIDIA’s latest Isaac Simulator, powered by their world-leading Omniverse platform. To start competing participants only require a few terminal commands and Python coding to leverage this powerful ecosystem.
Submissions for the challenge are through our challenge page on EvalAI. Please review the submission guidelines and instructions there before entering. Winning teams will be notified and invited to share their work in the Embodied AI workshop at CVPR 2022.
We look forward to seeing the new research and ideas that competing teams will produce.
- February 14th – Challenge launch
- May 31st – Close of challenge, and paper submissions due
- June 19th-24th – CVPR 2022
*The RVSU challenge is supported by the QUT Centre for Robotics (QCR), and Australian Centre for Robotic Vision (ACRV).