Related references
Note: Only part of the references are listed.
Article
Computer Science, Artificial Intelligence
Kai Han et al.
Summary: Transformer, a deep neural network with a self-attention mechanism, has been initially used in natural language processing and is now gaining attention in computer vision tasks. Transformer-based models perform as well as or even better than convolutional and recurrent neural networks in various visual benchmarks. This paper reviews vision transformer models, categorizes them based on different tasks, and analyzes their advantages and disadvantages. The discussed categories include backbone network, high/mid-level vision, low-level vision, and video processing. Efficient methods for applying transformer in real device-based applications are also explored. The challenges and further research directions for vision transformers are discussed as well.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Engineering, Electrical & Electronic
Liangyi Cui et al.
Summary: This paper proposes an improved Swin Transformer model for segmenting dense urban buildings from remote sensing images with complex backgrounds. A convolutional block attention module is utilized to focus on significant features, and hierarchical feature maps are fused to enhance the feature extraction process. The effectiveness and superiority of the proposed method are validated through ablation experiments and comparative studies, achieving an improvement of 1.3% in mean intersection-over-union compared to the original model.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2023)
Article
Environmental Sciences
Wei Yuan et al.
Summary: Extracting building data from remote sensing images has become more accurate with the emergence of deep learning technology, especially with the use of CNNs and Transformers. A Lite Swin transformer is proposed to simplify the calculation number of transformers, while the LiteST-Net model combines the features extracted by the Lite Swin transformer and CNN to better integrate global and local features. Comparison experiments show that LiteST-Net outperforms other networks in terms of all metrics and predicted image accuracy.
Article
Geography, Physical
Hamidreza Hosseinpour et al.
Summary: This research proposes a cross-modal gated fusion network (CMGFNet) for extracting building footprints from high-resolution remote sensing images and DSMs data. CMGFNet utilizes separate encoders to extract features from RGB and DSM data and employs cross-modal and multi-level feature fusion methods. Experimental results demonstrate that CMGFNet outperforms other state-of-the-art models, and extensive ablation study confirms the efficacy of all significant elements.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2022)
Article
Geography, Physical
Haonan Guo et al.
Summary: This study proposes a novel boundary refinement network (CBR-Net) for accurately extracting building footprints from remote sensing imagery. The CBR-Net progressively refines building predictions in a coarse-to-fine manner, while enhancing the model's ability to perceive and refine building edges. Experimental results demonstrate that CBR-Net outperforms other algorithms on various building datasets.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Reenul Reedha et al.
Summary: This paper explores the potential of attention-based deep networks (ViT) in weed and crop recognition using drone systems. It demonstrates that ViT models outperform state-of-the-art models with small labeled training datasets, showing promise in a wide range of remote sensing image analysis tasks.
Article
Geography, Physical
Huabing Huang et al.
Summary: This study developed a new method to estimate building height for all of China, resulting in a high-accuracy building height map that improves upon existing products. The new building height map is of great significance for the management of urban areas and further studies of urban environments.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Xiao Xiao et al.
Summary: This paper proposes a shifted-window (swin) Transformer-based encoding booster for efficient extraction of building areas in remote sensing images. By integrating the encoding booster in a specially designed U-shaped network, the feature-level fusion of local and large-scale semantics is achieved. Experimental results demonstrate that the proposed method achieves higher accuracy in extracting buildings of different scales compared to state-of-the-art networks.
Article
Computer Science, Information Systems
Zhongyu Sun et al.
Summary: This paper proposes a Hybrid Multi-resolution and Transformer semantic extraction Network (HMRT) that can provide a global receptive field, overcome the limitations of existing methods on high-resolution remote sensing images, and enhance scene understanding ability.
ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION
(2022)
Article
Engineering, Electrical & Electronic
Yong Zhou et al.
Summary: The neural network-based remote sensing image change detection method proposed in this study addresses the challenges of imaging interference and class imbalance problems under high-resolution conditions. It uses the siamese strategy and multi-head self-attention mechanism to reduce imaging differences and exploit inter-temporal information. It also incorporates a learnable multi-part feature learning module to obtain more comprehensive features. The mixed loss function strategy ensures effective convergence and excludes negative sample interference.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Proceedings Paper
Computer Science, Information Systems
G. F. Angelis et al.
Summary: This study compares different Transformer-based semantic segmentation architectures to evaluate their predictive performance and computational efficiency in extracting building footprints from remote sensing imagery. Four new architectures are introduced and compared with existing baselines.
2022 IEEE INTERNATIONAL SYMPOSIUM ON MULTIMEDIA (ISM)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Xiaoyi Dong et al.
Summary: CSWin Transformer is an efficient and effective Transformer-based backbone for general-purpose vision tasks. It achieves competitive performance by using the Cross-Shaped Window self-attention mechanism, Locally-enhanced Positional Encoding, and a hierarchical structure.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2022)
Article
Geography, Physical
Shouji Du et al.
Summary: This study proposes a semantic segmentation method for VHR images by combining a deep learning semantic segmentation model and object-based image analysis, which aims to capture precise outlines of ground objects and explore context information, achieving competitive overall accuracies for Vaihingen and Potsdam datasets.
INTERNATIONAL JOURNAL OF DIGITAL EARTH
(2021)
Article
Geochemistry & Geophysics
Zhuang Jia et al.
Summary: This letter proposes a simple but innovative end-to-end deep U-net-based model for hyperspectral image classification, which directly takes the whole HSI as network input and outputs predicted classes for each pixel location. The combination of classification loss and spatial constraint loss in the training stage enhances the spatial continuity and consistency of the predicted results, showing promising performance compared to existing CNN-based methods.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2021)
Article
Computer Science, Hardware & Architecture
Muhammad Alam et al.
Summary: This paper explores the application of Convolutional Neural Networks (CNN) for semantic segmentation of remote sensing images and proposes two models, SegNet and U-net, with index pooling. By integrating these models, an algorithm is presented which can achieve better multi-target segmentation compared to using the two models individually.
MOBILE NETWORKS & APPLICATIONS
(2021)
Article
Environmental Sciences
Haonan Guo et al.
Summary: Building footprint information is crucial for understanding urban processes and achieving environmentally sustainable urbanization. Automatic methods are needed to update building contour databases regularly, overcoming limitations in supervised approaches and accurately depicting building boundaries.
REMOTE SENSING OF ENVIRONMENT
(2021)
Article
Environmental Sciences
Yakoub Bazi et al.
Summary: This paper proposes a remote-sensing scene-classification method based on vision transformers, which utilize multihead attention mechanisms to establish long-range contextual relationships between pixels in images. The approach involves dividing images into patches, converting them into sequences, and applying data augmentation techniques for improved classification performance. The study also demonstrates the efficacy of compressing the network by pruning half of the layers while maintaining competitive classification accuracies.
Article
Environmental Sciences
Keyan Chen et al.
Summary: This paper explores the potential of using transformers for efficient building extraction and designs an efficient dual-pathway transformer structure that achieves state-of-the-art accuracy on benchmark datasets.
Article
Environmental Sciences
Wei Yuan et al.
Summary: The proposed multi-scale adaptive segmentation network model based on Swin Transformer (MSST-Net) addresses the limitation of convolutional neural networks in capturing global features by utilizing the self-attention mechanism. By using Swin Transformer to encode input images, decoding feature maps of different levels separately, fusing with convolution, and adjusting channels with a 1 x 1 kernel for final prediction map generation, the network model improves evaluation metrics on a WHU building dataset. This model emphasizes global features for remote sensing segmentation.
Review
Computer Science, Artificial Intelligence
Xiaohui Yuan et al.
Summary: This paper reviews the application of deep learning methods for semantic segmentation of remote sensing imagery and identifies challenges in handling non-traditional data as well as small datasets. Researchers are still facing difficulties in developing and evaluating new deep learning methods for remote sensing analysis.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Review
Geography, Physical
Qiqi Zhu et al.
Summary: Road extraction in satellite imagery is crucial for governments to plan infrastructure development and mobilize relief efforts worldwide. Recent advancements in deep learning have shown dominance in accurately labeling road pixels from high-resolution satellite images.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2021)
Article
Environmental Sciences
De-Yue Chen et al.
Summary: With the progress of urbanization, the management issues in the wildland-urban interface have become more serious. Building research is crucial in this area, and methods for extracting building information can be obtained from high-resolution remote sensing images or relevant agencies.
Review
Computer Science, Artificial Intelligence
Zhaoyang Niu et al.
Summary: This paper provides an overview of state-of-the-art attention models and defines a unified model suitable for most attention structures. It describes in detail each step of the attention mechanism implemented in the model and classifies existing attention models based on four criteria. Additionally, it summarizes the use of attention mechanisms in network architectures and typical applications.
Review
Environmental Sciences
Saman Ghaffarian et al.
Summary: Machine learning, especially deep learning, has become a key method in computer vision and remote sensing image processing. Researchers are exploring the use of attention mechanisms to enhance the performance of deep learning methods in remote sensing applications.
Proceedings Paper
Computer Science, Artificial Intelligence
Puxuan Yu et al.
Summary: This paper introduces two novel retrieval-oriented pretraining tasks to improve the performance of cross-lingual retrieval and transfer. By utilizing section alignment in multilingual Wikipedia to construct distant supervision data, the pretraining of retrieval-oriented language models is supported.
PROCEEDINGS OF THE WORLD WIDE WEB CONFERENCE 2021 (WWW 2021)
(2021)
Article
Engineering, Electrical & Electronic
Shuxian Dong et al.
Summary: This article proposes a pixel cluster CNN and spectral-spatial fusion (SSF) algorithm for hyperspectral image classification with small training samples. Experimental results demonstrate that the proposed algorithm outperforms traditional CNN and other studied algorithms in cases of small training sets.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Environmental Sciences
Xin-Yi Tong et al.
REMOTE SENSING OF ENVIRONMENT
(2020)
Article
Remote Sensing
Jiayun Liu et al.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2020)
Review
Mathematical & Computational Biology
Grace W. Lindsay
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE
(2020)
Article
Engineering, Electrical & Electronic
Luofan Dong et al.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2020)
Article
Remote Sensing
Shunping Ji et al.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2019)
Review
Remote Sensing
Aaron E. Maxwell et al.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2018)
Article
Environmental Sciences
Xin Pan et al.
Review
Computer Science, Artificial Intelligence
Yanming Guo et al.
INTERNATIONAL JOURNAL OF MULTIMEDIA INFORMATION RETRIEVAL
(2018)
Article
Environmental Sciences
Martin Langkvist et al.
Article
Geochemistry & Geophysics
Jun Wang et al.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2015)
Article
Engineering, Electrical & Electronic
Yansheng Li et al.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2015)
Article
Environmental Sciences
Yuguo Qian et al.
Article
Computer Science, Artificial Intelligence
Jing Chen et al.
ARTIFICIAL INTELLIGENCE REVIEW
(2011)
Article
Engineering, Electrical & Electronic
Martino Pesaresi et al.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2011)
Article
Geochemistry & Geophysics
Beril Sirmacek et al.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2010)