2000 character limit reached
Video Swin Transformers for Egocentric Video Understanding @ Ego4D Challenges 2022
Published 22 Jul 2022 in cs.CV | (2207.11329v1)
Abstract: We implemented Video Swin Transformer as a base architecture for the tasks of Point-of-No-Return temporal localization and Object State Change Classification. Our method achieved competitive performance on both challenges.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.