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:20250110T023312Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241205T154300
DTEND;TZID=Asia/Tokyo:20241205T155400
UID:siggraphasia_SIGGRAPH Asia 2024_sess135_papers_446@linklings.com
SUMMARY:GFFE: G-buffer Free Frame Extrapolation for Low-latency Real-time 
 Rendering
DESCRIPTION:Technical Papers\n\nSongyin Wu (University of California Santa
  Barbara); Deepak Vembar, Anton Sochenov, and Selvakumar Panneer (Intel Co
 rporation); Sungye Kim (Intel (now AMD)); Anton Kaplanyan (Intel Corporati
 on); and Ling-Qi Yan (University of California Santa Barbara)\n\nReal-time
  rendering has been embracing ever-demanding effects, such as ray tracing.
  However, rendering such effects in high resolution and high frame rate re
 mains challenging. Frame extrapolation methods, which do not introduce add
 itional latency as opposed to frame interpolation methods such as DLSS 3 a
 nd FSR 3, boost the frame rate by generating future frames based on previo
 us frames. However, it is a more challenging task because of the lack of i
 nformation in the disocclusion regions and complex future motions, and rec
 ent methods also have a high engine integration cost due to requiring G-bu
 ffers as input. We propose a \emph{G-buffer free} frame extrapolation meth
 od, GFFE, with a novel heuristic framework and an efficient neural network
 , to plausibly generate new frames in real time without introducing additi
 onal latency. We analyze the motion of dynamic fragments and different typ
 es of disocclusions, and design the corresponding modules of the extrapola
 tion block to handle them. After that, a light-weight shading correction n
 etwork is used to correct shading and improve overall quality. GFFE achiev
 es comparable or better results than previous interpolation and G-buffer d
 ependent extrapolation methods, with more efficient performance and easier
  integration.\n\nRegistration Category: Full Access, Full Access Supporter
 \n\nLanguage Format: English Language\n\nSession Chair: Changjian Li (Univ
 ersity of Edinburgh)
URL:https://asia.siggraph.org/2024/program/?id=papers_446&sess=sess135
END:VEVENT
END:VCALENDAR