Article
Environmental Sciences
Yufa Xia, Xin Xu, Fangling Pu
Summary: Synthetic aperture radar (SAR) imagery change detection (CD) is a crucial and challenging task. With the development of deep learning technologies, many deep learning methods have been presented for SAR CD and they outperform traditional methods. In this study, a pyramidal convolutional block attention network (PCBA-Net) is proposed for SAR image CD, which achieves superior performance by effectively utilizing both context information and detail information.
Article
Environmental Sciences
Liwen Zhang, Wenhao Wei, Bo Qiu, Ali Luo, Mingru Zhang, Xiaotong Li
Summary: This paper proposes a novel deep convolutional neural network for cloud image segmentation. The network utilizes a multibranch asymmetric convolution module and attention mechanisms to capture more contextual information, adaptively adjust feature channel weights, and strengthen useful features for cloud image segmentation. Experimental results demonstrate that the proposed method outperforms state-of-the-art methods in cloud image segmentation.
Article
Geochemistry & Geophysics
Qian Guo, Haipeng Wang, Feng Xu
Summary: A new hybrid approach combining scattering information enhancement (SIE) and an attention pyramid network (APN) is proposed for aircraft detection in synthetic aperture radar (SAR) images. Experimental results demonstrate the effectiveness of the method with an average precision (AP) of 83.25%.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Environmental Sciences
Sheng Zheng, Xinhong Hao, Chaoning Zhang, Wen Zhou, Lefan Duan
Summary: This study conducts an empirical study on deep classification for low-resolution SAR images. The results show that lightweight techniques developed in the high-resolution natural image domain can also be effective in the low-resolution SAR domain. The parameters are significantly reduced while improving the classification performance.
Article
Geochemistry & Geophysics
Lifu Chen, Ru Luo, Jin Xing, Zhenhong Li, Zhihui Yuan, Xingmin Cai
Summary: In this study, a geospatial transformer framework is proposed for aircraft detection in large-scale synthetic aperture radar (SAR) images. The framework consists of image decomposition, multiscale geospatial contextual attention network (MGCAN), and result recomposition. Experimental results show that the proposed method outperforms traditional methods in terms of detection performance.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Xiayang Xiao, Hecheng Jia, Penghao Xiao, Haipeng Wang
Summary: Due to the unique imaging mechanism of synthetic aperture radar (SAR), automatic aircraft detection in SAR images remains a challenge. To address this, a novel approach called adaptive deformable network (ADN) combined with peak feature fusion (PFF) is proposed. Experimental results on the GaoFen-3 (GF3) dataset demonstrate the effectiveness of the PFF-ADN with state-of-the-art performance.
Article
Mathematics
Guojun Mao, Guanyi Liao, Hengliang Zhu, Bo Sun
Summary: This paper proposes a lightweight and effective attention mechanism called M3Att, which can be embedded into object detection networks to significantly improve their performance. M3Att utilizes grouped convolutional layers and fusion of channel and spatial attention to achieve effective results on challenging object detection tasks.
Article
Computer Science, Artificial Intelligence
Gang Sha, Junsheng Wu, Bin Yu
Summary: In this paper, a scheme for spinal fracture lesions segmentation based on U-net was proposed, which introduces an attention module and dilated convolution to achieve more accurate lesions segmentation. The attention module focuses on specific regions to improve the model's recognition of lesions, while dilated convolution increases the receptive field for more lesion feature information. Experimental results show that the proposed network outperforms U-net in lesions segmentation performance.
NEURAL PROCESSING LETTERS
(2021)
Article
Geochemistry & Geophysics
Yan Zhao, Lingjun Zhao, Zhong Liu, Dewen Hu, Gangyao Kuang, Li Liu
Summary: In this article, a single shot detector called attentional feature refinement and alignment network (AFRAN) is proposed for aircraft detection in SAR imagery. The method achieves competitive accuracy and speed by refining and aligning informative characteristics of aircraft through carefully designed components.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Ru Luo, Lifu Chen, Jin Xing, Zhihui Yuan, Siyu Tan, Xingmin Cai, Jielan Wang
Summary: The Efficient Bidirectional Path Aggregation Attention Network (EBPA2N) addresses the challenges in aircraft detection from SAR images by utilizing the IEPA and ERSA modules, improving detection accuracy by effectively extracting advanced semantic and spatial information to better capture multi-scale scattering features of aircraft.
Article
Geochemistry & Geophysics
Xiang Yu, Ying Qian, Zhe Geng, Xiaohua Huang, Qinglu Wang, Daiyin Zhu
Summary: Recently, convolutional neural networks (CNNs) have shown great potential in synthetic aperture radar (SAR) target recognition. SAR images possess granular details and texture features of various scales, which are not typically considered in traditional CNN models. This article proposes an efficient isotopic architecture deep CNN (DCNN) called EMC(2)A-Net, which utilizes two residual blocks with multiscale receptive fields (RFs) and a multiscale feature cross-channel attention module (EMC(2)A module) to capture contextual features and enhance feature fusion. Experimental results on the MSTAR dataset demonstrate that EMC(2)A-Net outperforms other available models of the same type and has a relatively lightweight network structure.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Computer Science, Hardware & Architecture
Zhao Qiu, Lin Yuan, Lihao Liu, Zheng Yuan, Tao Chen, Zihan Xiao
Summary: A Wasserstein generative adversarial network with dilated convolution and deformable convolution (DDC-WGAN) is proposed for image completion, and experiments show that it outperforms traditional methods.
JOURNAL OF CIRCUITS SYSTEMS AND COMPUTERS
(2022)
Article
Geochemistry & Geophysics
Jiacheng Shi, Wei Liu, Haoyu Shan, Erzhu Li, Xing Li, Lianpeng Zhang
Summary: Scene classification in remote sensing plays a crucial role, and the development of convolutional neural networks (CNNs) has greatly improved its accuracy. However, the complexity and small size of objects in high-resolution RS images make CNNs ineffective. To overcome this challenge, a novel multibranch fusion attention network (MBFANet) is proposed to improve feature extraction and generalization performance. Using a multibranch fusion attention module (MBFAM), the model is able to focus on key cues that are difficult to classify, leading to improved performance on RS scene classification datasets.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Geochemistry & Geophysics
Yuming Xiang, Niangang Jiao, Rui Liu, Feng Wang, Hongjian You, Xiaolan Qiu, Kun Fu
Summary: This paper proposes a geometry-aware SAR image registration method that extracts inherent orientation features and focuses on geometry-invariant areas for accurate matching. The proposed method is universal under various imaging conditions and has been proven effective through extensive experiments on dozens of SAR images.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Remote Sensing
Ye Yuan, Yanxia Wu, Chuheng Tang, Yan Fu, Yulei Wu, Yan Jiang, Yize Zhao
Summary: In this article, a self-calibrated dilated convolutional neural network (SAR-SCDCNN) is proposed for SAR image despeckling. The method effectively suppresses speckle noise and preserves detailed features by adaptively extracting feature weights, performing convolutions, and feature fusion.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Yan Zhao, Lingjun Zhao, Zhong Liu, Dewen Hu, Gangyao Kuang, Li Liu
Summary: In this article, a single shot detector called attentional feature refinement and alignment network (AFRAN) is proposed for aircraft detection in SAR imagery. The method achieves competitive accuracy and speed by refining and aligning informative characteristics of aircraft through carefully designed components.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Yan Zhao, Lingjun Zhao, Boli Xiong, Gangyao Kuang
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2020)