Robust deep alignment network with remote sensing knowledge graph for zero-shot and generalized zero-shot remote sensing image scene classification

Li, Yansheng; Kong, Deyu; Tan, Yihua; Chen, Ling

doi:10.1016/j.isprsjprs.2021.08.001

Cited by 106 publications

(26 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the case of only a few available training samples, such a transfer learning task is then identified as few-shot learning. In the past few years, few-shot learning has been intensively explored to address various ML tasks, for instance, image classification (Li et al, 2021;Rußwurm et al, 2020;Vinyals et al, 2016;Wang et al, 2019), semantic segmentation (Hu et al, 2019;Zhang et al, 2019), and object detection (Kang et al, 2019;Li et al, 2022;Wang, Huang, et al, 2020). Different to the extensive efforts of few-short learning methods on image classification tasks, few-shot object detection received relatively limited research attention, because such a task requires the model to not only distinguish the object types but also to localize and count the targets in the given scene using a limited number of training samples (shots).…”

Section: Rel Ated Workmentioning

confidence: 99%

Improving OpenStreetMap missing building detection using few‐shot transfer learning in sub‐Saharan Africa

Herfort

Lautenbach

et al. 2022

Transactions in GIS

View full text Add to dashboard Cite

show abstract

Section: Rel Ated Workmentioning

confidence: 99%

Improving OpenStreetMap missing building detection using few‐shot transfer learning in sub‐Saharan Africa

Herfort

Lautenbach

et al. 2022

Transactions in GIS

View full text Add to dashboard Cite

show abstract

“…Liu et al [16] used image pairs as inputs and enhanced the robustness of the network by learning the regularization term through metrics. Existing multi-branch methods usually compare different scales of a single image, or compare sample distributions between different image features [18]. However, the difference between two images is rarely directly exploited to improve the class discriminative ability.…”

Section: Cnn Cnnmentioning

confidence: 99%

Pairwise Comparison Network for Remote-Sensing Scene Classification

Zhang

Zheng²,

2022

IEEE Geosci. Remote Sensing Lett.

View full text Add to dashboard Cite

Remote sensing scene classification aims to assign a specific semantic label to a remote sensing image. Recently, convolutional neural networks have greatly improved the performance of remote sensing scene classification. However, some confused images may be easily recognized as the incorrect category, which generally degrade the performance. The differences between image pairs can be used to distinguish image categories. This paper proposed a pairwise comparison network, which contains two main steps: pairwise selection and pairwise representation. The proposed network first selects similar image pairs, and then represents the image pairs with pairwise representations. The self-representation is introduced to highlight the informative parts of each image itself, while the mutualrepresentation is proposed to capture the subtle differences between image pairs. Comprehensive experimental results on two challenging datasets (AID, NWPU-RESISC45) demonstrate the effectiveness of the proposed network. The codes are provided in https://github.com/spectralpublic/PCNet.git.

show abstract

“…To evaluate the performance of the proposed HHTL framework, we conduct experiments on four public RS scene data sets, including the UC Merced (UCM) Land Use data set [16], Aerial Image data set (AID) [1], NWPU-RESISC45 (NWPU) data set [2] and RSSDIVCS data set [66], [67].…”

Section: Experiments a Data Set Introductionmentioning

confidence: 99%

Homo–Heterogenous Transformer Learning Framework for RS Scene Classification

Tang

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Remote sensing (RS) scene classification plays an essential role in the RS community and has attracted increasing attention due to its wide applications. Recently, benefiting from the powerful feature learning capabilities of convolutional neural networks (CNNs), the accuracy of RS scene classification has significantly been improved. Although the existing CNNbased methods achieve excellent results, there is still room for improvement. First, the CNN-based methods are adept at capturing the global information from RS scenes. Still, the context relationships hidden in RS scenes cannot be thoroughly mined. Second, due to the specific structure, it is easy for normal CNNs to exploit the heterogenous information from RS scenes. Nevertheless, the homogenous information, which is also crucial to comprehensively understand complex contents within RS scenes, does not get the attention it deserves. Third, most CNNs focus on establishing the relationships between RS scenes and semantic labels. However, the similarities between them are not considered deeply, which are helpful to distinguish the intra-/inter-class samples. To overcome the limitations mentioned above, we propose a homo-heterogenous transformer learning (HHTL) framework for RS scene classification in this paper. First, a patch generation module (PGM) is designed to generate homogenous and heterogenous patches. Then, a dual-branch feature learning module (FLM) is proposed to mine homogenous and heterogenous information within RS scenes simultaneously. In FLM, based on vision transformer, not only the global information but also the local areas and their context information can be captured. Finally, we design a classification module, which consists of a fusion sub-module and a metric-learning module. It can integrate homo-heterogenous information and compact/separate samples from the same/different RS scene categories. Extensive experiments are conducted on four public RS scene data sets. The encouraging results demonstrate that our HHTL framework can outperform many state-of-the-art methods. Our source codes are available at https://github.com/TangXu-Group/Remote-Sensing-Images-Classification/tree/main/HHTL.

show abstract

Robust deep alignment network with remote sensing knowledge graph for zero-shot and generalized zero-shot remote sensing image scene classification

Cited by 106 publications

References 36 publications

Improving OpenStreetMap missing building detection using few‐shot transfer learning in sub‐Saharan Africa

Improving OpenStreetMap missing building detection using few‐shot transfer learning in sub‐Saharan Africa

Pairwise Comparison Network for Remote-Sensing Scene Classification

Homo–Heterogenous Transformer Learning Framework for RS Scene Classification

Contact Info

Product

Resources

About