Skip to main content

ADT Object Detection 2023-4 Challenges [CLOSED]

Overview​

The 2023 Aria Digital Twin (ADT) challenges aim to accelerate research into 3D object detection and spatialization, serving as a catalyst for the broader research community.

Tasks​

Challenge One - 3D Object detection & tracking with scene generalization​

The ADT benchmark dataset consists of a number of moving objects and recorded sequences ground-truthed in two distinct scenes. For this task, generate accurate 6DoF poses for the same set of objects within recordings from a completely new and unseen scene we provide in the challenge data.

Challenge Two - Few-shot 3D Object detection & tracking​

We've provided a short period of ground-truth data for a specific target object within each sequence, your task is to leverage this data to accurately detect the same object and predict its 6DoF poses for the remaining duration of the sequence.

Prize per challenge​

First place: $10,000 and a certificate. Second and Third places will receive certificates and recognition on the Challenge leaderboard. Additionally, teams with the best performing or noteworthy submissions may be invited to present their work at the CVPR conference in June 2024.

Important dates​

  • Submission deadline for results: March 22, 2024 (11:59PM PST)
  • Presentation of awards: CVPR in June 2024 (pending confirmation of our workshop)

Key participation terms​

  • Entrants can register as an Individual or Team
  • Predictions are submitted online at Eval.ai
  • Winners agree to share a public report (e.g. whitepaper or publication) describing the method used for generating the Predictions
  • Participants are encouraged (but not required) to make source code used to generate the Predictions available under an Approved OSS License

Full terms can be found on ProjectAria.com/challenges/terms.

How to participate​

Follow the detailed instructions at Eval.ai.