QuadSampling: A Novel Sampling Method for Remote Implicit Neural 3D Reconstruction Based on Quad-Tree

Xu Qiang Hu, Yu Ping Wang*

*Corresponding author for this work

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

Abstract

Implicit neural representations have shown potential advantages in 3D reconstruction. But implicit neural 3D reconstruction methods require high-performance graphical computing power, which limits their application on low power consumption platforms. Remote 3D reconstruction framework can be employed to address this issue, but the sampling method needs to be further improved. We present a novel sampling method, QuadSampling, for remote implicit neural 3D reconstruction. By hierarchically sampling pixels within blocks with larger loss value, QuadSampling can result in larger average loss and help the neural learning process by better representing the shape of regions with different loss value. Thus, under the same amount of transmission, our QuadSampling can obtain more accurate and complete implicit neural representation of the scene. Extensive evaluations show that comparing with prior methods (i.e. random sampling and active sampling), our QuadSampling framework can improve the accuracy by up to 4%, and the completion ratio by about 1–2%.

Original languageEnglish
Title of host publicationComputer-Aided Design and Computer Graphics - 18th International Conference, CAD/Graphics 2023, Proceedings
EditorsShi-Min Hu, Yiyu Cai, Paul Rosin
PublisherSpringer Science and Business Media Deutschland GmbH
Pages314-328
Number of pages15
ISBN (Print)9789819996650
DOIs
Publication statusPublished - 2024
Event18th International Conference on Computer-Aided Design and Computer Graphics, CAD/Graphics 2023 - Shanghai, China
Duration: 19 Aug 202321 Aug 2023

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume14250 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference18th International Conference on Computer-Aided Design and Computer Graphics, CAD/Graphics 2023
Country/TerritoryChina
CityShanghai
Period19/08/2321/08/23

Keywords

  • Implicit Neural Representation
  • Remote 3D Reconstruction
  • Sampling Method

Fingerprint

Dive into the research topics of 'QuadSampling: A Novel Sampling Method for Remote Implicit Neural 3D Reconstruction Based on Quad-Tree'. Together they form a unique fingerprint.

Cite this