BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
X-LIC-LOCATION:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:19911015T033000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251218T030655Z
LOCATION:Meeting Room S423+S424\, Level 4
DTSTART;TZID=Asia/Hong_Kong:20251216T111200
DTEND;TZID=Asia/Hong_Kong:20251216T112300
UID:siggraphasia_SIGGRAPH Asia 2025_sess117_papers_1460@linklings.com
SUMMARY:Generating the Past, Present and Future from a Motion-Blurred Imag
 e
DESCRIPTION:SaiKiran Tedla (York University); Kelly Zhu (University of Tor
 onto, Vector Institute); Trevor Canham (York University); Felix Taubner (U
 niversity of Toronto, Vector Institute); Michael S. Brown (York University
 ); and Kiriakos N. Kutulakos and David B. Lindell (University of Toronto, 
 Vector Institute)\n\nWe seek to answer the question: what can a motion-blu
 rred image reveal about a scene's past, present, and future? Although moti
 on blur obscures image details and degrades visual quality, it also encode
 s information about scene and camera motion during an exposure. Previous t
 echniques leverage this information to estimate a sharp image from an inpu
 t blurry one, or to predict a sequence of video frames showing what might 
 have occurred at the moment of image capture. However, they rely on handcr
 afted priors or network architectures to resolve ambiguities in this inver
 se problem, and do not incorporate image and video priors on large-scale d
 atasets. As such, existing methods struggle to reproduce complex scene dyn
 amics and do not attempt to recover what occurred before or after an image
  was taken. Here, we introduce a new technique that repurposes a pre-train
 ed video diffusion model trained on internet-scale datasets to recover vid
 eos revealing complex scene dynamics during the moment of capture and what
  might have occurred immediately into the past or future. Our approach is 
 robust and versatile; it outperforms previous methods for this task, gener
 alizes to challenging in-the-wild images, and supports downstream tasks su
 ch as recovering camera trajectories, object motion, and dynamic 3D scene 
 structure.\n\nRegistration Category: Full Access, Full Access Supporter\n\
 nSession Chair: Qiang Fu (King Abdullah University of Science and Technolo
 gy (KAUST))\n\n
END:VEVENT
END:VCALENDAR