Research on the Graph Matching based on Shape Context and Sequential Monte Carlo

DOI

10.2991/amcce-15.2015.394How to use a DOI?

Keywords

Graph Matching; Shape Context; Sequential Monte Carlo

Abstract

According to the traditional shape context algorithm within the image distortion, excessive noise points and lower matching rate issues, we present a Sequential Monte Carlo algorithm based on graph matching. First, the image feature points were evenly distributed to compute shape context information. Second, remaining points obtain a histogram of all the feature points by the shape context information. Using histogram function to calculate the cost of the square distance cost, were begin to match. Finally, structuring the graph model, using graphical models construct the affinity matrix; the matrix will be close to integer quadratic programming use the Sequential Monte Carlo algorithm to find out the optimal matching schemes. Experimental results show that: the proposed algorithm in image matching to ensure a high rate, while images of different perspectives and images in quite different conditions has good robustness and stableness.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Meiju Liu
AU  - Hongyu Yang
AU  - Zhaohua Chen
AU  - Lingyan Li
PY  - 2015/04
DA  - 2015/04
TI  - Research on the Graph Matching based on Shape Context and Sequential Monte Carlo
BT  - Proceedings of the 2015 International Conference on Automation, Mechanical Control and Computational Engineering
PB  - Atlantis Press
SP  - 1175
EP  - 1180
SN  - 1951-6851
UR  - https://doi.org/10.2991/amcce-15.2015.394
DO  - 10.2991/amcce-15.2015.394
ID  - Liu2015/04
ER  -