Local interest region description using multiple support regions

Local image description is a key issue for local features related tasks in computer vision. A good descriptor should achieve high distinctiveness and robustness. In this paper, we propose a new descriptor called multi-support region improved weighted center symmetric local ternary pattern (MSR-IWCS-...

Full description

Saved in:
Bibliographic Details
Published inJournal of optics (New Delhi) Vol. 44; no. 3; pp. 290 - 297
Main Authors Huang, Mingming, Mu, Zhichun, Zeng, Hui
Format Journal Article
LanguageEnglish
Published New Delhi Springer India 01.09.2015
Subjects
Lasers
Optical Devices
Optics
Photonics
Physics
Physics and Astronomy
Research Article
Local ternary pattern
Affine transformation
Local image descriptor
Improved center symmetric local ternary pattern
Support region
Online AccessGet full text

Cover

More Information
Summary:Local image description is a key issue for local features related tasks in computer vision. A good descriptor should achieve high distinctiveness and robustness. In this paper, we propose a new descriptor called multi-support region improved weighted center symmetric local ternary pattern (MSR-IWCS-LTP). Unlike traditional descriptors, it uses multiple support regions that make the descriptor further enhance its discriminative ability. And by using a SIFT-like grid, the descriptor can contain more structural information. Extensive experimental results demonstrate the effectiveness of the proposed descriptor compared to existing state-of-the-art descriptors.
ISSN:0972-8821
0974-6900
DOI:10.1007/s12596-015-0257-6