BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240214T070242Z
LOCATION:Meeting Room C4.8\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231212T170000
DTEND;TZID=Australia/Melbourne:20231212T171000
UID:siggraphasia_SIGGRAPH Asia 2023_sess142_papers_644@linklings.com
SUMMARY:FuseSR: Super Resolution for Real-time Rendering through Efficient
  Multi-resolution Fusion
DESCRIPTION:Technical Communications, Technical Papers\n\nZhihua Zhong (St
 ate Key Lab of CAD&CG, Zhejiang University; Zhejiang University City Colle
 ge); Jingsen Zhu (State Key Lab of CAD&CG, Zhejiang University); Yuxin Dai
  (Zhejiang A&F University); Chuankun Zheng (State Key Lab of CAD&CG, Zheji
 ang University); Guanlin Chen (Zhejiang University City College); Yuchi Hu
 o (Zhejiang Lab; State Key Lab of CAD&CG, Zhejiang University); and Hujun 
 Bao and Rui Wang (State Key Lab of CAD&CG, Zhejiang University)\n\nThe wor
 kload of real-time rendering is steeply increasing as the demand for high 
 resolution, high refresh rates, and high realism rises, overwhelming most 
 graphics cards. To mitigate this problem, one of the most popular solution
 s is to render images at a low resolution to reduce rendering overhead, an
 d then manage to accurately upsample the low-resolution rendered image to 
 the target resolution, a.k.a. super-resolution techniques. Most existing m
 ethods focus on exploiting information from low-resolution inputs, such as
  historical frames. The absence of high frequency details in those LR inpu
 ts makes them hard to recover fine details in their high-resolution predic
 tions. In this paper, we propose an efficient and effective super-resoluti
 on method that predicts high-quality upsampled reconstructions utilizing l
 ow-cost high-resolution auxiliary G-Buffers as additional input. With LR i
 mages and HR G-buffers as input, the network requires to align and fuse fe
 atures at multi resolution levels. We introduce an efficient and effective
  H-Net architecture to solve this problem and significantly reduce renderi
 ng overhead without noticeable quality deterioration. Experiments show tha
 t our method is able to produce temporally consistent reconstructions in $
 4 \times 4$ and even challenging $8 \times 8$ upsampling cases at 4K resol
 ution with real-time performance, with substantially improved quality and 
 significant performance boost compared to existing works.\n\nRegistration 
 Category: Full Access\n\nSession Chair: Michael Gharbi (Adobe, MIT)
URL:https://asia.siggraph.org/2023/full-program?id=papers_644&sess=sess142
END:VEVENT
END:VCALENDAR