Related references
Note: Only part of the references are listed.
Article
Computer Science, Artificial Intelligence
Yuxiang Zhang et al.
Summary: This article introduces a Topological structure and Semantic information Transfer network (TSTnet) for the problem of cross-scene hyperspectral image classification. The method uses a graph structure to characterize topological relationships and utilizes graph convolutional networks (GCN) to analyze the data. Experimental results demonstrate that TSTnet outperforms other domain adaptation methods.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Geochemistry & Geophysics
Chao Tao et al.
Summary: This article analyzes and compares unsupervised feature learning, supervised feature learning, and self-supervised feature learning from the perspective of feature learning signals and provides a unified feature learning framework. It also discusses the advantages of self-supervised feature learning over the other two paradigms in remote sensing image understanding tasks and reviews existing self-supervised feature learning works in this field.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Mingyang Zhang et al.
Summary: This paper proposes a cross-domain self-taught network (CDSTN) for few-shot HSI classification. It combines domain adaptation (DA) and semi-supervised self-taught strategy to utilize labeled and unlabeled samples from source and target domains. Experimental results show that CDSTN achieves superior and stable performance with limited labeled samples.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Computer Science, Artificial Intelligence
Yuxiang Zhang et al.
Summary: Currently, there is an increasing attention on cross-scene hyperspectral image (HSI) classification. This study proposes a Single-source Domain Expansion Network (SDEnet) based on domain generalization to extend the domain of a model trained only on the source domain (SD) to the target domain (TD). The method uses generative adversarial learning and a specific generator architecture to generate extended domain (ED) with variable spatial and spectral information. Supervised contrastive learning and adversarial training are employed to optimize the discriminator and the generator, respectively, for better domain invariant representation. Experimental results demonstrate the superiority of the proposed method compared to state-of-the-art techniques. The source code is available at: https://github.com/YuxiangZhang-BIT/IEEE_TIP_SDEnet.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2023)
Article
Geochemistry & Geophysics
Qingjie Zeng et al.
Summary: IDLN is an iterative distribution learning network designed for few-shot remote sensing image scene classification, consisting of three modules: similarity distribution learning, label mapping, and attention-based feature calibration, to enhance classification performance.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Gong Cheng et al.
Summary: SPNet is a lightweight and effective few-shot image classification model, utilizing prototype self-calibration and intercalibration methods to generate more accurate prototypes and predictions through optimizing three loss functions, demonstrating competitive performance compared with other advanced few-shot image classification approaches.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Wang Miao et al.
Summary: The paper proposes a semi-supervised representation consistency Siamese network (SS-RCSN) for remote-sensing image scene classification, utilizing GAN to extract discriminative features and reducing the differences between labeled and unlabeled data, experimental results demonstrate superior performance compared to other semi-supervised learning methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Xiaomin Li et al.
Summary: This study introduces a self-supervised contrastive learning framework for few-shot remote sensing scene classification, which effectively addresses the issue of classification with limited data and accelerates model convergence through an end-to-end training approach. Additionally, an attention module is incorporated to fuse multi-scale spatial features from classification targets of different sizes.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Computer Science, Information Systems
Donghang Yu et al.
Summary: This paper proposes an end-to-end feature fusion method that aggregates features from dual paths (AFDP) to extract discriminative features from remote sensing images. Experimental results show that the AFDP method outperforms current state-of-the-art methods in remote sensing image classification.
Article
Geography, Physical
Qingjie Zeng et al.
Summary: In this article, a few-shot scene classification model for remote sensing images is proposed. The model utilizes self-attention and mutual-attention modules to learn feature correlations and optimizes the deep network using a task-specific contrastive loss function. Experimental results demonstrate the superiority of the proposed method in terms of accuracy compared to other related methods.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Xu Tang et al.
Summary: This article introduces a remote sensing scene classification algorithm that can be trained with a few labeled samples. By introducing a dual-branch network and a class-level prototype guided learning algorithm, it can capture abundant land covers with multiscale and multiangle information in remote sensing scenes. The experiments demonstrate that the algorithm performs well with small labeled data.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Jiangtao Xie et al.
Summary: This paper proposes a deep Brownian Distance Covariance (DeepBDC) method for few-shot classification. The central idea is to learn image representations by measuring the discrepancy between joint characteristic functions of embedded features and product of marginals. Extensive evaluations show that DeepBDC significantly outperforms other methods and achieves state-of-the-art results.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2022)
Article
Geochemistry & Geophysics
Jie Geng et al.
Summary: A feature enhanced superpixel hypergraph neural network (FESHNN) is proposed for PolSAR image classification in this article, aiming to take full advantage of spatial and polarimetric features. Experimental results on three PolSAR datasets demonstrate its superior classification performance compared with other approaches.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Baoquan Zhang et al.
Summary: Few-shot remote sensing scene classification is an important task that aims to recognize novel scene classes with limited examples. Existing methods have not fully utilized the unique characteristics of remote sensing images, so we propose a novel meta-learning framework called SGMNet, which utilizes scene graph matching to address this issue. Experimental results show that our method achieves superior performance on several datasets.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Hong Ji et al.
Summary: This article proposes a two-stage framework for few-shot scene classification of optical remote sensing images. The first stage learns a general-purpose representation and two auxiliary objectives. The second stage aims to find an expected prototype with the minimal distance to all samples within the same class.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Computer Science, Artificial Intelligence
Yuxiang Zhang et al.
Summary: This study proposes a new domain adaptation method that combines few-shot learning with domain alignment based on graph information aggregation to address the issue of reduced classification performance in the presence of new classes in target data. By training on source data with all label samples and target data with a few label samples, two key blocks are designed to extract and compare features and distributions across different domains, and cross-domain graph alignment methods are used to mitigate the impact of domain shift on few-shot learning. Experimental results demonstrate the superiority of the proposed method.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Geochemistry & Geophysics
Hong Ji et al.
Summary: Small data has big potential for artificial intelligence. Few-shot learning is a promising approach that aims to efficiently learn a model capable of recognizing novel classes with limited training samples. This study focuses on few-shot scene classification of optical remote sensing images and proposes a method that combines different objectives and regularization techniques to enhance the generalization ability. Experimental results show that the proposed method outperforms state-of-the-art approaches on public remote sensing benchmarks.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Tengfei Gong et al.
Summary: This research aims to achieve novel category recognition for remote sensing images with a few shot samples. By combining meta-learning and self-supervised learning, it utilizes a pretrained network and unlabeled target domain images for few-shot classification.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Yuxiang Zhang et al.
Summary: The study introduces a new domain adaptive method, discriminative cooperative alignment (DCA) of subspace and distribution, to address the challenge of cross-scene HSI classification. Experimental results demonstrate that the DCA method significantly outperforms other domain-adaptive approaches.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Agriculture, Multidisciplinary
Wadii Boulila et al.
Summary: This study proposes a novel distributed deep learning-based approach for classifying big remote sensing images. Experiments conducted on a dataset covering multiple regions of Saudi Arabia demonstrate high classification accuracy and faster speed compared to other state-of-the-art classification methods.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2021)
Article
Geochemistry & Geophysics
Lingjun Li et al.
Summary: In this article, we propose an end-to-end network called discriminative learning of adaptive match network (DLA-MatchNet) to enhance few-shot remote sensing image scene classification. By utilizing attention technique and adaptive matching, our method automatically discovers discriminative regions and selects semantically relevant sample pairs, respectively. Experimental results show the effectiveness of our model in tackling the challenges of few-shot scene classification in remote sensing images.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Environmental Sciences
Qingjie Zeng et al.
Summary: A method of prototype calibration with a feature-generating model is proposed for few-shot remote sensing image scene classification to address the issue of prototype feature expression deviation. Experiments demonstrate that the proposed method can achieve superior classification accuracies on NWPU-RESISC45 and WHU-RS19 datasets.
Article
Environmental Sciences
Joseph Kim et al.
Summary: In this paper, a multi-scale feature fusion network for few-shot remote sensing scene classification is proposed by integrating a novel self-attention feature selection module, denoted as SAFFNet. The feature weighting value can be fine-tuned by the support set in the few-shot learning task, significantly improving the accuracy of few-shot classification.
Article
Environmental Sciences
Yong Li et al.
Summary: The study introduces a novel model, Meta-FSEO, to enhance the performance of few-shot remote sensing scene classification in varying urban scenes. By deploying self-supervised embedding optimization and adaptive generalization in new tasks, the model demonstrates great generalization capability in remote sensing scene classification among different cities.
Proceedings Paper
Computer Science, Artificial Intelligence
Mingkai Zheng et al.
Summary: This paper introduces a weakly supervised contrastive learning framework that improves the quality of self-supervised representations across different datasets and achieves state-of-the-art results in semi-supervised learning. Using two projection heads, with one head performing instance discrimination task and the other generating weak labels based on a graph-based method, the framework demonstrates superior performance compared to existing methods, even outperforming SimCLRv2 in certain scenarios.
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021)
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Kai Huang et al.
Summary: The pseudo-loss confidence metric (PLCM) measures data credibility by the loss distribution of pseudo-labels and integrates various tasks, effectively selecting and strengthening the classifier.
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021)
(2021)
Article
Engineering, Electrical & Electronic
Mohsen Ansari et al.
Summary: A convolutional kernel classifier (CKC) was proposed for hyperspectral remote sensing images to address the computational complexity and end-to-end learning paradigm issues. The CKC uses Nystrom approximation and deep architecture to learn the optimal combination of basis kernels, outperforming all competitor methods in classifying four benchmark hyperspectral datasets. The runtime of CKC is also lower than MKL algorithms counterparts, indicating its efficiency in high-dimensional kernel approximation and classification accuracy improvement.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Computer Science, Artificial Intelligence
Si-Wei Chen et al.
Summary: This study examines the impact of different speckle reduction methods on SAR ship detection performances, finding that the influences of different speckle filters can be positive or negative. A speckle-free SAR ship detection approach is proposed, utilizing the SPN indicator for target identification, achieving good detection performance.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)
Article
Geochemistry & Geophysics
Jie Geng et al.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2020)
Article
Environmental Sciences
John W. Jones
Article
Geography, Physical
Wenzhi Zhao et al.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2019)
Article
Geography, Physical
Wei Han et al.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2018)
Article
Computer Science, Artificial Intelligence
Mostafa Mehdipour Ghazi et al.
Article
Engineering, Electrical & Electronic
Gong Cheng et al.
PROCEEDINGS OF THE IEEE
(2017)
Article
Remote Sensing
Gong Cheng et al.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2013)
Article
Geochemistry & Geophysics
Jie Feng et al.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2011)