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 B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241205T092800
DTEND;TZID=Asia/Tokyo:20241205T094200
UID:siggraphasia_SIGGRAPH Asia 2024_sess125_papers_206@linklings.com
SUMMARY:Filtering-Based Reconstruction for Gradient-Domain Rendering
DESCRIPTION:Technical Papers\n\nDifei Yan and Shaokun Zheng (Tsinghua Univ
 ersity), Ling-Qi Yan (University of California Santa Barbara), and Kun Xu 
 (Tsinghua University)\n\nGradient-domain rendering methods reconstruct col
 or images based on the Poisson equation with gradients from correlated sam
 pling. The relatively low variance in the gradient estimation facilitates 
 convergence but the inevitable noises make the solving process prone to un
 pleasant spiky artifacts.\n\nWe present a gradient-guided filtering approa
 ch for reconstruction, which avoids the instability from the direct usage 
 of noisy gradients. Instead, we model the output color of each pixel as a 
 weighted combination of neighboring pixels, where the gradients are used a
 s guidance to compute optimized filtering weights. The gradients are enhan
 ced before being used in gradient-guided filtering. A coarse-to-fine strat
 egy is also employed to make use of information from a larger scale. \n\nE
 xperiments demonstrate that our method achieves the best reconstruction re
 sults for gradient-domain renderings compared to existing techniques. Besi
 des, our method has two desirable properties: first, our method is not lea
 rning-based so it does not require an extra training step and would be mor
 e robust for unseen scenes; second, our method is designed to be asymptoti
 c unbiased.\n\nRegistration Category: Full Access, Full Access Supporter\n
 \nLanguage Format: English Language\n\nSession Chair: Wenzel Jakob (École 
 Polytechnique Fédérale de Lausanne)
URL:https://asia.siggraph.org/2024/program/?id=papers_206&sess=sess125
END:VEVENT
END:VCALENDAR