Presentation

<
Streetscapes: Large-scale Consistent Street View Generation Using Autoregressive Video Diffusion
SessionConsistent Text-to-Image
DescriptionOur method generates Streetscapes — long sequences of views through a synthesized city-scale scene. We build on video diffusion models, but in an autoregressive framework that easily scales to long camera trajectories. We train our system on the unique Google Street View data, allowing controlling generations by scene layouts and camera poses.
Authors
Boyang Deng
Stanford University
Google Research
Richard Tucker
Google Research
Zhengqi Li
Google Research
Leonidas Guibas
Stanford University
Google Research
Noah Snavely
Google Research
Cornell University
Gordon Wetzstein
Stanford University
Event Type
Technical Paper
TimeMonday, 29 July 20242:30pm - 2:40pm MDT
LocationMile High 1
ACM Digital Library Journal Papers' PDFs
Conference Papers' PDFs
Session Time & Location
Sunday, 28 July 20246:00pm - 8:45pm MDTBluebird Ballroom
Monday, 29 July 20242:00pm - 3:30pm MDTMile High 1
Interest Areas
Research & Education
Recordings
Livestreamed
Recorded
Keywords
AI
Machine Learning
Rendering
Registration Categories
Full Conference
Full Conference Supporter
Virtual Access
Exhibitor Full Conference
Monday
Next PresentationNext PresentationDMHomo: Learning Homography with Diffusion Models