From Unordered Range Images to 3D Models: A Fully Automatic Multiview Correspondence Algorithm

Authors:
Affiliations:
Venue:
TPCG '04 Proceedings of the Theory and Practice of Computer Graphics 2004 (TPCG'04)
Year:
2004

Citing 0
Cited 3

A Novel Representation and Feature Matching Algorithm for Automatic Pairwise Registration of Range Images

International Journal of Computer Vision
Three-Dimensional Model-Based Object Recognition and Segmentation in Cluttered Scenes

IEEE Transactions on Pattern Analysis and Machine Intelligence
3D Geometric Scale Variability in Range Images: Features and Descriptors

International Journal of Computer Vision

Quantified Score

Hi-index	0.01

Visualization

Abstract

In this paper, we present a fully automatic multiviewcorrespondence algorithm based on our tensor representation.Our algorithm can automatically establish correspondencesbetween a set of unordered range images of an objectand register them in a common coordinate basis. Existingtechniques of correspondence are either manual or inefficientand require the ordering of range images accordingto their overlap. Our algorithm does not suffer from anysuch limitation. We also present a complete and fully automatic3D modeling framework by integrating our multiviewcorrespondence algorithm with a global registration, an integrationand a surface reconstruction algorithm. Our resultsshow that our algorithm is fully automatic, applicableto free-form objects, efficient and independent of the resolutionand sampling of the range images.