Computer Perception

Design and evaluation of full computer perception systems.

Workshop at ECCV, 23 October 2022

Overview

Much of the research in Computer Vision has focused on solving individual tasks like image or video classification, object detection, object tracking, to name only a few. These settings are suitable for narrow applications, but for complex applications like embodied intelligent assistants or self-driving cars we will need full perception models that deal with multiple modalities and have integrated scene understanding and reasoning capabilities comparable to humans.

In this workshop, we propose to zoom out from Computer Vision and discuss more generally about Computer Perception, with leading thinkers from Computer Vision/Machine Learning and Cognitive Sciences.

We will cover both modelling challenges and evaluation best practices.

For evaluation, among others, we will discuss the recently released Perception Test, that you can try yourself here.

Schedule

The program features invited talks (including Q&As) and a panel discussion.

All times listed are in Israel local time.

Slides presenting the Perception Test available here

12:00 - 12:15 Opening notes - Joao Carreira
12:15 - 13:10 Perception Test (part 1) - Viorica Patraucean, Dima Damen
13:10 - 14:00 Lunch break
14:00 - 14:35 Keynote Olga Russakovsky
14:35 - 15:30 Perception Test (part 2) - Ankush Gupta, Yi Yang, Carl Doersch, Adria Recasens Continente, Antoine Miech, Viorica Patraucean
15:30 - 16:00 Coffee break
16:00 - 16:35 Keynote Aude Oliva
16:35 - 17:10 Keynote Matt Botvinick
17:10 - 17:45 Keynote Michael Auli
17:45 - 18:00 Coffee break
18:00 - 18:35 Keynote Daniel Yamins
18:35 - 19:10 Keynote Jitendra Malik
19:10 - 19:55 Panel discussion
19:55 - 20:00 Closing notes

Organizers

Viorica Pătrăucean

Deepmind
Research Scientist at DeepMind.
Research: computer vision, scalable learning, biologically plausible learning.

Joao Carreira

Deepmind
Research Scientist at DeepMind.
Research: video processing, general perception systems.

Dima Damen

Bristol University
Professor of Computer Vision.
Research: computer vision, video understanding, perception benchmarks.

Andrew Zisserman

University of Oxford
Professor of Computer Vision Engineering at Oxford and a Royal Society Research Professor.
Research: computer vision, machine learning.