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 325-AB\, Level 3\, West Wing
DTSTART;TZID=Asia/Seoul:20221207T140000
DTEND;TZID=Asia/Seoul:20221207T153000
UID:siggraphasia_SIGGRAPH Asia 2022_sess162_papers_146@linklings.com
SUMMARY:Rhythmic Gesticulator: Rhythm-Aware Co-Speech Gesture Synthesis wi
 th Hierarchical Neural Embeddings
DESCRIPTION:Technical Communications, Technical Papers\n\nRhythmic Gesticu
 lator: Rhythm-Aware Co-Speech Gesture Synthesis with Hierarchical Neural E
 mbeddings\n\nAo, Gao, Lou, Chen, Liu\n\nAutomatic synthesis of realistic c
 o-speech gestures is an increasingly important yet challenging task in art
 ificial embodied agent creation. Previous systems mainly focus on generati
 ng gestures in an end-to-end manner, which leads to difficulties in mining
  the clear rhythm and semantics due to the complex yet subtle harmony betw
 een speech and gestures. We present a novel co-speech gesture synthesis me
 thod that achieves convincing results both on the rhythm and semantics. Fo
 r the rhythm, our system contains a robust rhythm-based segmentation pipel
 ine to ensure the temporal coherence between the vocalization and gestures
  explicitly. For the gesture semantics, we devise a mechanism to effective
 ly disentangle both low- and high-level neural embeddings of speech and mo
 tion based on linguistic theory. The high-level embedding corresponds to s
 emantics, while the low-level embedding relates to subtle variations. Last
 ly, we build correspondence between the hierarchical embeddings of the spe
 ech and the motion, resulting in rhythm- and semantics-aware gesture synth
 esis. Evaluations with existing objective metrics, a newly proposed rhythm
 ic metric, and human feedback show that our method outperforms state-of-th
 e-art systems by a clear margin.\n\nRegistration Category: FULL ACCESS, ON
 -DEMAND ACCESS\n\nLanguage: ENGLISH\n\nFormat: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_146&sess=sess16
 2
END:VEVENT
END:VCALENDAR