Article ID Journal Published Year Pages File Type
536302 Pattern Recognition Letters 2015 7 Pages PDF
Abstract

•We propose a novel operator called orthogonal symmetric local ternary pattern.•A descriptor called weighted orthogonal symmetric local ternary pattern is proposed.•The OS-LTP variance of the local region is applied to construct the descriptor.•Proposed approach demonstrates the high effectiveness and efficiency.

Local image description is a key issue for local features related tasks in computer vision. A good descriptor should achieve two competing goals which are high-quality description and low computational complexity. In this paper, we propose a novel operator called orthogonal symmetric local ternary pattern (OS-LTP), achieving robustness against noise interference and discriminative ability for describing texture structure. Then, based on the proposed OS-LTP operator, we introduce a new descriptor, named weighted orthogonal symmetric local ternary pattern (WOS-LTP). Unlike traditional descriptors, the WOS-LTP descriptor is constructed by using the OS-LTP variance of the local region as an adaptive weight to adjust the contribution of the OS-LTP code in histogram calculation. Extensive experimental results demonstrate the effectiveness and efficiency of the new descriptor compared with existing state-of-the-art descriptors.

Graphical abstractLocal image description is a key issue for local features related tasks in computer vision. A good descriptor should achieve two competing goals which are high-quality description and low computational complexity. In this paper, we propose a novel operator called orthogonal symmetric local ternary pattern (OS-LTP), achieving robustness against noise interference and discriminative ability for red describing texture structure. Then, based on the proposed OS-LTP operator, we introduce a new descriptor, named weighted orthogonal symmetric local ternary pattern (WOS-LTP). Unlike traditional descriptors, the WOS-LTP descriptor is constructed by using the OS-LTP variance of the local region as an adaptive weight to adjust the contribution of the OS-LTP code in histogram calculation. Extensive experimental results demonstrate the effectiveness and efficiency of the new descriptor compared with existing state-of-the-art descriptors.Figure optionsDownload full-size imageDownload high-quality image (173 K)Download as PowerPoint slide

Related Topics
Physical Sciences and Engineering Computer Science Computer Vision and Pattern Recognition
Authors
, , , ,