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:20250110T023311Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241203T153100
DTEND;TZID=Asia/Tokyo:20241203T154300
UID:siggraphasia_SIGGRAPH Asia 2024_sess108_papers_373@linklings.com
SUMMARY:LVCD: Reference-based Lineart Video Colorization with Diffusion Mo
 dels
DESCRIPTION:Technical Papers\n\nZhitong Huang (City University of Hong Kon
 g); Mohan Zhang (Wechat, Tencent Inc.); and Jing Liao (City University of 
 Hong Kong)\n\nWe propose the first video diffusion framework for reference
 -based lineart video colorization. Unlike previous works that rely solely 
 on image generative models to colorize lineart frame by frame, our approac
 h leverages a large-scale pretrained video diffusion model to generate col
 orized animation videos. This approach leads to more temporally consistent
  results and is better equipped to handle large motions. Firstly, we intro
 duce Sketch-guided ControlNet which provides additional control to finetun
 e an image-to-video diffusion model for controllable video synthesis, enab
 ling the generation of animation videos conditioned on lineart. We then pr
 opose Reference Attention to facilitate the transfer of colors from the re
 ference frame to other frames containing fast and expansive motions. Final
 ly, we present a novel scheme for sequential sampling, incorporating the O
 verlapped Blending Module and Prev-Reference Attention, to extend the vide
 o diffusion model beyond its original fixed-length limitation for long vid
 eo colorization. Both qualitative and quantitative results demonstrate tha
 t our method significantly outperforms state-of-the-art techniques in term
 s of frame and video quality, as well as temporal consistency. Moreover, o
 ur method is capable of generating high-quality, long temporal-consistent 
 animation videos with large motions, which is not achievable in previous w
 orks. Our code and model are available at https://luckyhzt.github.io/lvcd.
 \n\nRegistration Category: Full Access, Full Access Supporter\n\nLanguage 
 Format: English Language\n\nSession Chair: I-Chao Shen (The University of 
 Tokyo)
URL:https://asia.siggraph.org/2024/program/?id=papers_373&sess=sess108
END:VEVENT
END:VCALENDAR