BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260817T171530Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241205T153100
DTEND;TZID=Asia/Tokyo:20241205T154300
UID:siggraphasia_SIGGRAPH Asia 2024_sess134_papers_860@linklings.com
SUMMARY:VidPanos: Generative Panoramic Videos from Casual Panning Videos
DESCRIPTION:Jingwei Ma (University of Washington); Erika Lu, Roni Paiss, a
 nd Shiran Zada (Google Deepmind); Aleksander Holynski (University of Calif
 ornia Berkeley, Google Deepmind); Tali Dekel (Weizmann Institute of Scienc
 e, Google Deepmind); Brian Curless (University of Washington, Google Deepm
 ind); and Michael Rubinstein and Forrester Cole (Google Deepmind)\n\nStitc
 hing frames of a panning video into a panoramic photograph is a well-under
 stood problem for stationary scenes. When objects are moving, however, a s
 till panorama is not enough to capture the scene. \nWe present a method fo
 r synthesizing a panoramic video from a casually-captured panning video, a
 s if the original video were captured with a wide-angle camera. We pose pa
 norama synthesis as a space-time outpainting problem, where we aim to crea
 te a full panoramic video of the same length as the input video. Consisten
 t completion of the space-time volume requires a powerful, realistic prior
  over video content and motion, for which we adapt generative video models
 . Existing generative models do not, however, immediately extend to panora
 ma completion, as we show. We instead apply video generation as a componen
 t of our panorama synthesis system, and demonstrate how to exploit the str
 engths of the models while minimizing their limitations. Our system can cr
 eate video panoramas for a range of in-the-wild scenes including people, v
 ehicles, and flowing water, as well as stationary background features.\n\n
 Registration Category: Full Access, Full Access Supporter\n\nLanguage Form
 at: English Language\n\nSession Chair: Nanxuan Zhao (Adobe Research)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_860&sess=sess134
END:VEVENT
END:VCALENDAR