Parallel alignment of a large number of range images

T. Oishi, A. Nakazawa, R. Sagawa, R. Kurazume, Katsushi Ikeuchi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

24 Citations (Scopus)

Abstract

We describe a method for parallel alignment of multiple range images. It is difficult to align a large number of range images simultaneously. Therefore, we developed the parallel method to improve the time and memory performances of the alignment process. Although a general simultaneous alignment algorithm searches correspondences for all pairs of all range images by rejecting redundant dependencies, our method makes it possible to accelerate computation time and reduce the amount of memory used. Since the computation between two range images can be preformed independently, each correspondence pair of range images is assigned to each node. Because the computation time is proportional to the number of vertices assigned to each node, by assigning the pairs so that the number of vertices computed is equal on each node, the load on each node is effectively distributed. The heuristic algorithms for graph partitioning are applied to this problem in order to reduce the amount of memory used on each node. The method was tested on a 16-processor PC cluster, where it demonstrated the high extendibility and the performance improvement in time and memory.

Original languageEnglish
Title of host publicationProceedings - 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003
PublisherIEEE Computer Society
Pages195-202
Number of pages8
ISBN (Electronic)0769519911
DOIs
Publication statusPublished - 2003
Event4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003 - Banff, Canada
Duration: Oct 6 2003Oct 10 2003

Publication series

NameProceedings of International Conference on 3-D Digital Imaging and Modeling, 3DIM
Volume2003-January
ISSN (Print)1550-6185

Other

Other4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003
CountryCanada
CityBanff
Period10/6/0310/10/03

All Science Journal Classification (ASJC) codes

  • Computer Graphics and Computer-Aided Design
  • Computer Science Applications
  • Computer Vision and Pattern Recognition

Fingerprint Dive into the research topics of 'Parallel alignment of a large number of range images'. Together they form a unique fingerprint.

Cite this