A Composite Descriptor for Shape Image Retrieval

DOI

10.2991/amcce-15.2015.135How to use a DOI?

Keywords

retrieval; special feature; shape matrix Fourier descriptor(SMFD); Histogram of Distances; Euclidean distance

Abstract

In order to quickly retrieval the interested image from the mass images, this paper proposed a composite descriptor based on contour shape and spacial feature for image retrieval. Firstly, we extract the contour by the shape matrix Fourier descriptor(SMFD).Then, use the adjacent contour sampling points and image center to form a triangle, the triangle have a only Circumcircle , use the radius of the Circumcircle to form a distance histograms, by adjusting the weight , combined the two methods, measures the similarity among the images with Euclidean distance. Finally, the experiment verified the feasibility and efficiency of the method. Experimental results show that the method is simple and fast , it has the character of transfer, rotation and scale invariant, it can achieve better retrieval accuracy and recall rates than the other method.

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  - Xinjian Wang
AU  - Guangchun Luo
AU  - Ke Qin
PY  - 2015/04
DA  - 2015/04
TI  - A Composite Descriptor for Shape Image Retrieval
BT  - Proceedings of the 2015 International Conference on Automation, Mechanical Control and Computational Engineering
PB  - Atlantis Press
SP  - 1225
EP  - 1230
SN  - 1951-6851
UR  - https://doi.org/10.2991/amcce-15.2015.135
DO  - 10.2991/amcce-15.2015.135
ID  - Wang2015/04
ER  -