1 paper
Chuhan Zhang, Ankush Gupta, Andrew Zisserman
The objective of this work is to learn an object-centric video representation, with the aim of improving transferability to novel tasks, i.e., tasks different from the pre-training…