FusionMLS

Highly dynamic 3D reconstruction with consumer-grade RGB-D cameras

Siim Meerits, Diego Thomas, Vincent Nozick, Hideo Saito

Research output: Contribution to journalArticle

Abstract

Multi-view dynamic three-dimensional reconstruction has typically required the use of custom shutter-synchronized camera rigs in order to capture scenes containing rapid movements or complex topology changes. In this paper, we demonstrate that multiple unsynchronized low-cost RGB-D cameras can be used for the same purpose. To alleviate issues caused by unsynchronized shutters, we propose a novel depth frame interpolation technique that allows synchronized data capture from highly dynamic 3D scenes. To manage the resulting huge number of input depth images, we also introduce an efficient moving least squares-based volumetric reconstruction method that generates triangle meshes of the scene. Our approach does not store the reconstruction volume in memory, making it memory-efficient and scalable to large scenes. Our implementation is completely GPU based and works in real time. The results shown herein, obtained with real data, demonstrate the effectiveness of our proposed method and its advantages compared to stateof- the-art approaches.

Original languageEnglish
JournalComputational Visual Media
DOIs
Publication statusAccepted/In press - 2018 Jan 1

Fingerprint

Camera shutters
Cameras
Data storage equipment
Data acquisition
Interpolation
Topology
Costs
Graphics processing unit

Keywords

  • 3D reconstruction
  • GPU
  • motion capture
  • RGB-D cameras

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

FusionMLS : Highly dynamic 3D reconstruction with consumer-grade RGB-D cameras. / Meerits, Siim; Thomas, Diego; Nozick, Vincent; Saito, Hideo.

In: Computational Visual Media, 01.01.2018.

Research output: Contribution to journalArticle

@article{76e1b6aa6ba245dbaf74a2e43118bfbb,
title = "FusionMLS: Highly dynamic 3D reconstruction with consumer-grade RGB-D cameras",
abstract = "Multi-view dynamic three-dimensional reconstruction has typically required the use of custom shutter-synchronized camera rigs in order to capture scenes containing rapid movements or complex topology changes. In this paper, we demonstrate that multiple unsynchronized low-cost RGB-D cameras can be used for the same purpose. To alleviate issues caused by unsynchronized shutters, we propose a novel depth frame interpolation technique that allows synchronized data capture from highly dynamic 3D scenes. To manage the resulting huge number of input depth images, we also introduce an efficient moving least squares-based volumetric reconstruction method that generates triangle meshes of the scene. Our approach does not store the reconstruction volume in memory, making it memory-efficient and scalable to large scenes. Our implementation is completely GPU based and works in real time. The results shown herein, obtained with real data, demonstrate the effectiveness of our proposed method and its advantages compared to stateof- the-art approaches.",
keywords = "3D reconstruction, GPU, motion capture, RGB-D cameras",
author = "Siim Meerits and Diego Thomas and Vincent Nozick and Hideo Saito",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/s41095-018-0121-0",
language = "English",
journal = "Computational Visual Media",
issn = "2096-0433",
publisher = "Tsinghua University Press",

}

TY - JOUR

T1 - FusionMLS

T2 - Highly dynamic 3D reconstruction with consumer-grade RGB-D cameras

AU - Meerits, Siim

AU - Thomas, Diego

AU - Nozick, Vincent

AU - Saito, Hideo

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Multi-view dynamic three-dimensional reconstruction has typically required the use of custom shutter-synchronized camera rigs in order to capture scenes containing rapid movements or complex topology changes. In this paper, we demonstrate that multiple unsynchronized low-cost RGB-D cameras can be used for the same purpose. To alleviate issues caused by unsynchronized shutters, we propose a novel depth frame interpolation technique that allows synchronized data capture from highly dynamic 3D scenes. To manage the resulting huge number of input depth images, we also introduce an efficient moving least squares-based volumetric reconstruction method that generates triangle meshes of the scene. Our approach does not store the reconstruction volume in memory, making it memory-efficient and scalable to large scenes. Our implementation is completely GPU based and works in real time. The results shown herein, obtained with real data, demonstrate the effectiveness of our proposed method and its advantages compared to stateof- the-art approaches.

AB - Multi-view dynamic three-dimensional reconstruction has typically required the use of custom shutter-synchronized camera rigs in order to capture scenes containing rapid movements or complex topology changes. In this paper, we demonstrate that multiple unsynchronized low-cost RGB-D cameras can be used for the same purpose. To alleviate issues caused by unsynchronized shutters, we propose a novel depth frame interpolation technique that allows synchronized data capture from highly dynamic 3D scenes. To manage the resulting huge number of input depth images, we also introduce an efficient moving least squares-based volumetric reconstruction method that generates triangle meshes of the scene. Our approach does not store the reconstruction volume in memory, making it memory-efficient and scalable to large scenes. Our implementation is completely GPU based and works in real time. The results shown herein, obtained with real data, demonstrate the effectiveness of our proposed method and its advantages compared to stateof- the-art approaches.

KW - 3D reconstruction

KW - GPU

KW - motion capture

KW - RGB-D cameras

UR - http://www.scopus.com/inward/record.url?scp=85052307362&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85052307362&partnerID=8YFLogxK

U2 - 10.1007/s41095-018-0121-0

DO - 10.1007/s41095-018-0121-0

M3 - Article

JO - Computational Visual Media

JF - Computational Visual Media

SN - 2096-0433

ER -