This workshop will consider challenges and opportunities at the frontier of 3D perception of scenes and objects directly from single cameras, i.e., images and videos. There will be a number of keynote talks as well as two debates aimed at discussing recent trends, with emphasis on the role of foundation models in monocular 3D reconstruction and scene understanding. Our goal is to disseminate state-of-the-art technical knowledge on 3D perception, assess where we are, and to provide a venue for meaningful and vigorous debate of current challenges and ways to overcome them.

In this third edition of the Frontiers of Monocular 3D Perception workshop, we will focus on the role of foundation models in 3D perception, and in particular how geometry can be used as a principled way to learn and apply such models. We believe geometry, as well as other physically grounded inductive priors, will play a key role in the generation of embodied foundation models, that are able to interact with the physical world and learn from these additional constraints in ways that purely visual models cannot. Thus, we hope to inspire and guide further research in that direction, including both well-established and up-and-coming groups.

We aim to discuss recent exciting progress on 3D tasks using a single RGB camera, including:

• monocular depth estimation

• 3D detection

• shape modeling

• scene reconstruction

• visual odometry

Program Details (GMT+2) 

• 1:30pm  -    1:40pm — Introductory remarks

• 1:40pm  -    2:10pm — Amir Zamir

• 2:10pm  -    2:40pm — Noah Snavely  

• 2:40pm  -    3:10pm — Angela Dai

• 3:10pm  -    3:45pm — Coffee Break

• 3:45pm  -    4:15pm — Rana Hanocka

• 4:15pm  -    4:45pm — Carl Vondrick

• 4:45pm  -    5:15pm — Oisin Mac Aodha

• 5:15pm  -    5:55pm — Panel: Katerina Fragkidaki, Leo Guibas, Jaime Spencer Martin, Haochen Wang

• 5:55pm  -    6:00pm — Concluding remarks

Speakers

Panelists

Organizers

The workshop is mainly organized by researchers from Toyota Technological Institute at Chicago (TTIC) & Toyota Research Institute (TRI).

Contact

If you have any questions about the workshop, please contact us at  vitor.guizilini@tri.global / igor.vasiljevic@tri.global.