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:20240214T070247Z
LOCATION:Meeting Room C4.8\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231214T121500
DTEND;TZID=Australia/Melbourne:20231214T123000
UID:siggraphasia_SIGGRAPH Asia 2023_sess150_papers_491@linklings.com
SUMMARY:GeoLatent: A Geometric Approach to Latent Space Design for Deforma
 ble Shape Generators
DESCRIPTION:Technical Papers\n\nHaitao Yang, Bo Sun, Liyan Chen, Amy Pavel
 , and Qixing Huang (University of Texas at Austin)\n\nWe study how to opti
 mize the latent space of neural shape generators that map latent codes to 
 3D deformable shapes. The key focus is to look at a deformable shape gener
 ator from a differential geometry perspective. We define a Riemannian metr
 ic based on as-rigid-as-possible and as-conformal-as-possible deformation 
 energies. Under this metric, we study two desired properties of the latent
  space: 1) straight-line interpolations in latent codes follow geodesic cu
 rves; 2) latent codes disentanglement pose and shape variations at differe
 nt scales. Strictly enforcing the geometric interpolation property, howeve
 r, only applies if the metric matrix is a constant. We show how to achieve
  this property approximately by enforcing that geodesic extrapolations are
  axis-aligned, i.e., extrapolations along coordinate axis follow geodesic 
 curves. In addition, we introduce a novel approach that decouples pose and
  shape variations via generalized eigendecomposition. We also study effici
 ent regularization terms for learning deformable shape generators, e.g., t
 hat promote smooth interpolations. Experimental results on benchmark datas
 ets show that our approach leads to interpretable latent codes, improves t
 he generalizability of synthetic shapes, and enhances performance in geode
 sic interpolation, geodesic shooting, and parallel translation application
 s.\n\nRegistration Category: Full Access\n\nSession Chair: Peng-Shuai Wang
  (Peking University)
URL:https://asia.siggraph.org/2023/full-program?id=papers_491&sess=sess150
END:VEVENT
END:VCALENDAR