Skip to main navigation Skip to search Skip to main content

A Three-Dimensional Inverse Rendering Method Based on Point Light Source Position Estimation

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Inverse rendering, a technique for reconstructingthree-dimensional (3D) geometric structures and surface materials from two-dimensional (2D) images, is of great significance in fields such as robotic manipulation, virtual reality, and augmented reality. However, existing inverse rendering methods based on neural implicit representations are primarily suitable for scenarios with fixed light sources and struggle with dynamic point lighting. This paper introduces a novel 3D inverse rendering method that estimates the positions of point light sources, facilitating the 3D reconstruction of geometrical structures and the determination of light source locations from a collection of images captured from various perspectives, even under dynamic illumination conditions from an unknown point light source.

Original languageEnglish
Title of host publicationASIG 2024 - Proceedings of the 2nd Asia Symposium on Image and Graphics
PublisherAssociation for Computing Machinery, Inc
Pages171-176
Number of pages6
ISBN (Electronic)9798400709906
DOIs
StatePublished - 26 Apr 2025
Event2nd Asia Symposium on Image and Graphics, ASIG 2024 - Sanya, China
Duration: 20 Dec 202422 Dec 2024

Publication series

NameASIG 2024 - Proceedings of the 2nd Asia Symposium on Image and Graphics

Conference

Conference2nd Asia Symposium on Image and Graphics, ASIG 2024
Country/TerritoryChina
CitySanya
Period20/12/2422/12/24

Keywords

  • 3D reconstruction
  • computer vision
  • neural implicit representation

Fingerprint

Dive into the research topics of 'A Three-Dimensional Inverse Rendering Method Based on Point Light Source Position Estimation'. Together they form a unique fingerprint.

Cite this