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:20250110T023312Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241204T113100
DTEND;TZID=Asia/Tokyo:20241204T114300
UID:siggraphasia_SIGGRAPH Asia 2024_sess113_papers_847@linklings.com
SUMMARY:Neural Light Spheres for Implicit Image Stitching and View Synthes
 is
DESCRIPTION:Technical Papers\n\nIlya Chugunov and Amogh Joshi (Princeton U
 niversity), Kiran Murthy and Francois Bleibel (Google Inc.), and Felix Hei
 de (Princeton University)\n\nChallenging to capture, and challenging to di
 splay on a cellphone screen, the panorama paradoxically remains both a sta
 ple and underused feature of modern mobile camera applications. In this wo
 rk we address both of these challenges with a spherical neural light field
  model for implicit panoramic image stitching and re-rendering; able to ac
 commodate for depth parallax, view-dependent lighting, and local scene mot
 ion and color changes during capture. Fit during test-time to an arbitrary
  path panoramic video capture -- vertical, horizontal, random-walk -- thes
 e neural light spheres jointly estimate the camera path and a high-resolut
 ion scene reconstruction to produce novel wide field-of-view projections o
 f the environment. Our single-layer model avoids expensive volumetric samp
 ling, and decomposes the scene into compact view-dependent ray offset and 
 color components, with a total model size of 80 MB per scene, and real-tim
 e (50 FPS) rendering at 1080p resolution. We demonstrate improved reconstr
 uction quality over traditional image stitching and radiance field methods
 , with significantly higher tolerance to scene motion and non-ideal captur
 e settings.\n\nRegistration Category: Full Access, Full Access Supporter\n
 \nLanguage Format: English Language\n\nSession Chair: Forrester Cole (Goog
 le)
URL:https://asia.siggraph.org/2024/program/?id=papers_847&sess=sess113
END:VEVENT
END:VCALENDAR