Presentation

ASSET: Autoregressive Semantic Scene Editing With Transformers at High Resolutions
Event Type
Technical Paper
Presentation Types
In Person
Interest Areas
Research & Education
Recordings
This session WILL NOT be recorded.
Registration Categories
Full Conference Supporter
Full Conference
Exhibitor Additional Full Conference
Exhibitor Full Conference
Keywords
Imaging/Video
TimeTuesday, 9 August 20229:41am - 9:49am PDT
LocationEast Building, Ballroom A/B
DescriptionWe present ASSET, a transformer architecture able to produce state-of-the-art results in image editing at high resolutions. The key idea of ASSET is a novel attention mechanism that efficiently captures long-range interactions important for realistic image synthesis. We present qualitative and quantitative results, along with user studies, demonstrating ASSET's effectiveness.