The 4th Perception Test Challenge at ECCV
Do large multimodal models truly understand the spatial structure of the world they perceive? Given that they are largely trained on passive, internet-scale data, how do they represent environments across scales, from the tabletop in a tea preparation video to the city-scale layout of an hour-long walking tour?
Put your model to test and win prizes totalling 20K EUR!
NEW this year: We run a unified videoQA track probing spatial intelligence in table-top and city-scale videos. A single model has to handle both scales to be considered eligible.
Challenge
Speakers
University College London
Technical University Munich
Cornell University
Workshop Agenda
- 09:00 - 09:10 Welcome and introduction
- 09:10 - 09:30 Challenge overview (Viorica Patraucean)
- 09:30 - 10:00 Keynote (Laura Leal-Taixé)
- 10:00 - 10:45 Coffee break
- 10:45 - 11:15 Keynote (Noah Snavely)
- 11:15 - 12:00 Challenge results
- 12:00 - 12:30 Keynote (Hugo Spiers)
- 12:30 - 13:00 Panel discussion
Challenge Timeline
- June 20: Challenge server with validation tasks goes live
- July 1: Test data and tasks released
- August 18: Deadline for submissions
- August 20: Deadline for tech report submissions
- September 2: Decision to participants
Organizers
Viorica Patraucean
Google DeepMind
Joao Carreira
Google DeepMind
Dima Damen
Bristol University
Andrew Zisserman
Oxford University