Presentation
Be Decisive: Noise-Induced Layouts for Multi-Subject Generation
SessionDeep Image Editing
DescriptionText-to-image diffusion models struggle with multi-subject generation due to subject leakage. Prior methods impose external layouts that conflict with the model’s prior, harming alignment and natural composition. We introduce a method that leverages the layout encoded in the initial noise, promoting alignment and natural compositions while preserving the model’s diversity.
Event Type
Technical Paper
TimeWednesday, 13 August 202510:45am - 10:55am PDT
LocationWest Building, Rooms 118-120
ACM Digital Library
Journal Papers' PDFs
Conference Papers' PDFs
Conference Papers' PDFs
Session Time & Location
Sunday, 10 August 20256:00pm - 8:45pm PDTWest Building, Ballroom AB
Wednesday, 13 August 202510:45am - 12:35pm PDTWest Building, Rooms 118-120
Livestreamed
Recorded
