Discovering Novel Actions from Open World Egocentric Videos with Object-Grounded Visual Commonsense Reasoning.

Sanjoy Kundu, Shubham Trehan, Sathyanarayanan N. Aakur

Browse the full ECCV paper archive.