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:20250110T023313Z
LOCATION:Hall B5 (1)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241206T134200
DTEND;TZID=Asia/Tokyo:20241206T135600
UID:siggraphasia_SIGGRAPH Asia 2024_sess145_papers_960@linklings.com
SUMMARY:SING: Stability-Incorporated Neighborhood Graph
DESCRIPTION:Technical Papers\n\nDiana Marin (Technical University of Vienn
 a); Amal Dev Parakkat (LTCI, Telecom ParisTech; Institut Polytechnique de 
 Paris); Stefan Ohrhallinger and Michael Wimmer (Technical University of Vi
 enna); Steve Oudot (Institut national de recherche en informatique et en a
 utomatique (INRIA)); and Pooran Memari (CNRS, Inria, LIX; École Polytechni
 que, IP-Paris)\n\nWe introduce the Stability-Incorporated Neighborhood Gra
 ph (SING), a novel density-aware structure designed to capture the intrins
 ic geometric properties of a point set. We improve upon the spheres-of-inf
 luence graph by incorporating additional features to offer more flexibilit
 y and control in encoding proximity information and capturing local densit
 y variations. Through persistence analysis on our proximity graph, we prop
 ose a new clustering technique and explore additional variants incorporati
 ng extra features for the proximity criterion. Alongside the detailed anal
 ysis and comparison to evaluate its performance on various datasets, our e
 xperiments demonstrate that the proposed method can effectively extract me
 aningful clusters from diverse datasets with variations in density and cor
 relation. Our application scenarios underscore the advantages of the propo
 sed graph over classical neighborhood graphs, particularly in terms of par
 ameter tuning.\n\nRegistration Category: Full Access, Full Access Supporte
 r\n\nLanguage Format: English Language\n\nSession Chair: Hao (Richard) Zha
 ng (Simon Fraser University, Amazon)
URL:https://asia.siggraph.org/2024/program/?id=papers_960&sess=sess145
END:VEVENT
END:VCALENDAR