.png){kind=icon}
返回
Exploring joint encoding of multi-direction local binary patterns for image classification
.png){kind=icon}
PRE
AI
DOI：10.1007/s11042-017-5319-0.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Local binary pattern (LBP) has been investigated as prominent feature in image classification for decades because of its strong discriminative ability and simple computation complexity, however, it extracts patterns in local circular space, ignoring the patterns located in local line-geometry space. In this paper, we propose a novel feature named joint encoding of multi-direction LBP (JEMDLBP) for image representation. Concretely, due to direction information are highly valuable to reflect the fundamental properties of images, we firstly develop direction LBP (DLBP) codes in local line space of four directions, including 0 degrees, 45 degrees, 90 degrees and 135 degrees directions. Secondly, the DLBP dominant patterns are generated by employing statistical analysis based on the occurrence rates of DLBP codes, which has high adaptive power to database and direction. Thirdly, we employ joint encoding strategy to capture co-occurrence patterns between adjacent directions with the hope that stronger local line structure can be extracted. Finally, to validate the effectiveness and efficiency of JEMDLBP, extensive experiments are carried out on eight databases (four face recognition, one face expression recognition, one palmprint recognition, two texture classification). Excellent results show that JEMDLBP achieves a performance that is competitive or better than several related LBP-like descriptors.
Keyword:
Image classification
Feature representation
Direction local binary patterns
Co-occurrence
Joint encoding
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
3
论文数:
2.0W
被引数:
3.2W
机构
引用论文
The FERET database and evaluation procedure for face-recognition algorithms人脸识别算法的FERET数据库和评估程序
PREAI
分享收藏Crystalline‐State Reaction with Allosteric Effect in Spin‐Crossover, Interpenetrated Networks with Magnetic and Optical Bistability具有磁和光学双稳态的自旋交叉，互穿网络中具有变构效应的晶态反应
PREAI
分享收藏Arsenic cluster superlattice in gallium arsenide grown by low-temperature molecular-beam epitaxy砷化镓中由低温分子束外延生长的砷团簇超晶格
PREAI
分享收藏Description of interest regions with local binary patterns用局部二值模式描述感兴趣区域.png)
PATTERN RECOGNITION
PATTERN RECOGNITIONIF7.6
PREAI
分享收藏分享收藏Nanocrystalline silicon films as multifunctional material for optoelectronic and photovoltaic applications纳米硅薄膜作为光电和光伏应用的多功能材料
PREAI
分享收藏
查看更多内容