BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Seoul
X-LIC-LOCATION:Asia/Seoul
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:KST
DTSTART:18871231T000000
DTSTART:19881009T020000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230103T035309Z
LOCATION:Room 324\, Level 3\, West Wing
DTSTART;TZID=Asia/Seoul:20221207T140000
DTEND;TZID=Asia/Seoul:20221207T153000
UID:siggraphasia_SIGGRAPH Asia 2022_sess159_papers_157@linklings.com
SUMMARY:VToonify: Controllable High-Resolution Portrait Video Style Transf
 er
DESCRIPTION:Technical Communications, Technical Papers\n\nVToonify: Contro
 llable High-Resolution Portrait Video Style Transfer\n\nYang, Jiang, Liu, 
 Loy\n\nGenerating high-quality artistic portrait videos is an important an
 d desirable task in computer graphics and vision. Although a series of suc
 cessful portrait image toonification models built upon the powerful StyleG
 AN have been proposed, these image-oriented methods have obvious limitatio
 ns when applied to videos, such as the fixed frame size, the requirement o
 f face alignment, missing non-facial details and temporal inconsistency. I
 n this work, we investigate the challenging controllable high-resolution p
 ortrait video style transfer by introducing a novel VToonify framework. Sp
 ecifically, VToonify leverages the mid- and high-resolution layers of Styl
 eGAN to render high-quality artistic portraits based on the multi-scale co
 ntent features extracted by an encoder to better preserve the frame detail
 s. The resulting fully convolutional architecture accepts non-aligned face
 s in videos of variable size as input, contributing to complete face regio
 ns with natural motions in the output. Our framework is compatible with ex
 isting StyleGAN-based image toonification models to extend them to video t
 oonification, and inherits appealing features of these models for flexible
  style control on color and intensity. This work presents two instantiatio
 ns of VToonify built upon Toonify and DualStyleGAN for collection-based an
 d exemplar-based portrait video style transfer, respectively. Extensive ex
 perimental results demonstrate the effectiveness of our proposed VToonify 
 framework over existing methods in generating high-quality and temporally-
 coherent artistic portrait videos with flexible style controls.\n\nRegistr
 ation Category: FULL ACCESS, ON-DEMAND ACCESS\n\nLanguage: ENGLISH\n\nForm
 at: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_157&sess=sess15
 9
END:VEVENT
END:VCALENDAR