Skip to content

Adaptive evolution strategy sample consensus for 3D reconstruction from two cameras

Research output: Contribution to journalArticlepeer-review

  • Yuichiro Toda
  • Hsu Horng Yz
  • Takayuki Matsuno
  • Mamoru Minami
  • Dr Dalin Zhou
RANdom SAmple Consensus (RANSAC) has been applied to many 3D image processing problems such as homography matrix estimation problems and shape detection from 3D point clouds, and is one of the most popular robust estimator methods. However, RANSAC has a problem related to the trade-off between computational cost and stability of search because RANSAC is based on random sampling. In our previous work, we proposed Adaptive Evolution Strategy SAmple Consensus (A-ESSAC) as a new robust estimator, and we applied ESSAC to the homography matrix estimation for 3D SLAM using RGB-D camera. A-ESSAC is based on Evolution Strategy to maintain the genetic diversity. Furthermore, ESSAC has two heuristic searches. One is a search range control for reducing the computational cost of RANSAC. The other is adaptive/self-adaptive mutation for changing the search strategy of A-ESSAC according to the best fitness value. In this paper, we apply A-ESSAC to 3D reconstruction method using two cameras, and we show an experimental result, and discuss the effectiveness of the proposed method.
Original languageEnglish
Number of pages9
JournalArtificial Life and Robotics
Early online date24 Apr 2020
DOIs
Publication statusEarly online - 24 Apr 2020

Documents

  • Adaptive Evolution Strategy SAmple Consensus for 3D Reconstruction from Two Cameras

    Rights statement: This is a post-peer-review, pre-copyedit version of an article published in Artificial Life and Robotics. The final authenticated version is available online at: http://dx.doi.org/10.1007/s10015-020-00603-9.

    Accepted author manuscript (Post-print), 3.86 MB, PDF document

    Due to publisher’s copyright restrictions, this document is not freely available to download from this website until: 24/04/21

Related information

Relations Get citation (various referencing formats)

ID: 19706077