.png){kind=icon}
返回
Grouped Subspace Linear Semantic Alignment for Hyperspectral Image Transfer Learning
.png){kind=icon}
PRE
AI
DOI：10.1109/TGRS.2022.3184691.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Transfer learning (TL) offers an effective way to reduce the demand for labeled samples in remote sensing image classification. However, existing TL methods have some limitations. Simple linear TL methods cannot align the source and TARget domains (TARs) well when the data shift is complicated, while nonlinear methods consist of many learnable parameters and often need many labeled samples. To overcome these issues, we design a novel grouped subspace linear semantic alignment (G-SLSA) algorithm, which consists of four main ingredients. First, inspired by the linear supervised transfer learning (LSTL) approach, we propose subspace linear semantic alignment (SLSA) aiming to reduce the demand for labeled samples. Second, considering the heterogeneity of class-level data shift, we extend SLSA to G-SLSA through a grouped alignment strategy, which can reduce the data shift by decomposing a multiclass TL task into a set of binary subtasks. Third, considering the demand of subtasks fusion on posterior probabilities, we propose a robust posterior probability estimation method for the binary generalized learning vector quantization (GLVQ) that is used in G-SLSA. Finally, a pairwise coupling (PWC) method is applied to fuse the results of each subtask. Experimental results conducted on three popular hyperspectral datasets demonstrate that G-SLSA outperforms other traditional and state-of-the-art deep learning (DL) methods, with an overall accuracy (OA) of 80.78 +/- 4.28% for Pavia City-University of Pavia (PC-UP) dataset TL scenario when five samples per class are available in the TAR.
Keyword:
Transfer learning
Adaptation models
Training
Task analysis
Semantics
Hyperspectral imaging
Mathematical models
Generalized learning vector quantization (GLVQ)
hyperspectral image (HSI) classification
subspace alignment (SA)
transfer learning (TL)
期刊
IF:
8.6
论文数:
2.1W
被引数:
10.7W
机构
引用论文
分享收藏
Deep support vector machine for hyperspectral image classification基于深度支持向量机的高光谱图像分类.png)
PATTERN RECOGNITION
PATTERN RECOGNITIONIF7.6
PREAI
分享收藏Advances in Hyperspectral Image and Signal Processing A comprehensive overview of the state of the art高光谱图像和信号处理的进展综述
PREAI
分享收藏Density and Thermal Conductivity Measurements for Silicon Melt by Electromagnetic Levitation under a Static Magnetic Field在静态磁场下通过电磁悬浮测量硅熔体的密度和热导率
PREAI
分享收藏
查看更多内容