Article
Geochemistry & Geophysics
Jianing Wang, Siying Guo, Runhu Huang, Linhao Li, Xiangrong Zhang, Licheng Jiao
Summary: The proposed DcCapsGAN effectively addresses the challenges in hyperspectral image classification and significantly improves the accuracy and performance of classification.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Qingyu Hou, Zhipeng Wang, Fanjiao Tan, Ye Zhao, Haoliang Zheng, Wei Zhang
Summary: This letter proposes a robust infrared small target detection network based on deep learning, which can detect small targets of different sizes and low SNRs in complex backgrounds, and has better effectiveness and robustness against existing algorithms.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Environmental Sciences
Alireza Arabameri, Subodh Chandra Pal, Fatemeh Rezaie, Rabin Chakrabortty, Asish Saha, Thomas Blaschke, Mariano Di Napoli, Omid Ghorbanzadeh, Phuong Thao Thi Ngo
Summary: This paper explores the predictive capacity of different approaches to landslide susceptibility modeling using artificial intelligence. The results show that the CDT-Multiboost model is the excellent model with high accuracy and is effective for improving spatial prediction of landslide susceptibility.
GEOCARTO INTERNATIONAL
(2022)
Article
Geochemistry & Geophysics
Konrad Heidler, Lichao Mou, Celia Baumhoer, Andreas Dietz, Xiao Xiang Zhu
Summary: This study proposes a new model that combines deep learning methods for segmenting and delineating coastlines. By combining building blocks from different frameworks and using deep supervision and hierarchical attention mechanism, the training effectiveness is improved. The advantages of this approach over traditional methods and other deep learning methods are demonstrated on a challenging dataset of Antarctic coastlines.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Gong Cheng, Xuxiang Sun, Ke Li, Lei Guo, Junwei Han
Summary: The article introduced an effective defense framework PSGAN for RSI scene classification, which trains the classifier by generating examples to combat known and unknown attacks. Experimental results demonstrated the great effectiveness of PSGAN.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Lingjun Li, Xiwen Yao, Gong Cheng, Mingliang Xu, Jungong Han, Junwei Han
Summary: This article aims to achieve one-shot object detection by mimicking human one-shot learning ability, proposing a solo-to-collaborative dual-attention network (SCoDANet) to enhance image feature representations hierarchically. The method effectively solves the issues of extracting related information between query and target images and intractable feature extraction of query class in target images.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Danfeng Hong, Lianru Gao, Renlong Hang, Bing Zhang, Jocelyn Chanussot
Summary: This paper introduces a multimodal deep learning model, EndNet, for hyperspectral and LiDAR data classification, which enhances material identification ability by fusing features and reconstructing multimodal inputs.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Zhiyong Lv, Fengjun Wang, Guoqing Cui, Jon Atli Benediktsson, Tao Lei, Weiwei Sun
Summary: Land cover change detection using remote sensing images is important for various applications. This article proposes a novel neural network method with spatial-spectral attention mechanism and multiscale dilation convolution modules to enhance the detection accuracy. The proposed method achieves superior performance compared to state-of-the-art methods in terms of quantitative evaluation metrics and visual performance.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Leiyao Liao, Lan Du, Yuchen Guo
Summary: In the field of remote sensing image processing, a method called FDDA is proposed to improve semi-supervised SAR target detection by utilizing a decoding module and a domain-adaptation module. Experimental results show promising performance in SAR target detection with limited labeled SAR images using this approach.
Article
Geochemistry & Geophysics
Gong Cheng, Liming Cai, Chunbo Lang, Xiwen Yao, Jinyong Chen, Lei Guo, Junwei Han
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
Gaodian Zhou, Weitao Chen, Qianshan Gui, Xianju Li, Lizhe Wang
Summary: The article introduces a method to improve the accuracy of road extraction from high-resolution remote-sensing images using a split depth-wise separable graph convolutional network. The results of the experiment show that this method performs better in extracting covered and tiny roads.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Nico Lang, Nikolai Kalischek, John Armston, Konrad Schindler, Ralph Dubayah, Jan Dirk Wegner
Summary: NASA's GEDI mission aims to advance understanding of forests' role in the global carbon cycle. A novel supervised machine learning approach has been developed to interpret GEDI waveforms and regress canopy top height globally with reliable predictive uncertainty estimates.
REMOTE SENSING OF ENVIRONMENT
(2022)
Article
Geochemistry & Geophysics
Yi Xiao, Xin Su, Qiangqiang Yuan, Denghong Liu, Huanfeng Shen, Liangpei Zhang
Summary: A novel fusion strategy of temporal grouping projection and an accurate alignment module is proposed for satellite video super-resolution (VSR) in this article. By enhancing the alignment and fusion methods, the spatial resolution and dynamic analysis of satellite videos are effectively improved. Extensive experiments demonstrate that this method outperforms current state-of-the-art VSR methods on Jilin-1 satellite video.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Yibin Ren, Xiaofeng Li, Xiaofeng Yang, Huan Xu
Summary: This study developed a deep learning model, DAU-Net, to classify sea ice and open water from SAR images with high accuracy. By integrating the dual-attention mechanism into U-Net, improvements in pixel-level classification and IoU were achieved. Experimental results showed that DAU-Net outperformed the original U-Net and DenseNetFCN models.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Computer Science, Interdisciplinary Applications
Onat Dalmaz, Mahmut Yurt, Tolga Cukur
Summary: This paper proposes a novel generative adversarial approach, ResViT, which combines the contextual sensitivity of vision transformers, the precision of convolution operators, and the realism of adversarial learning. Demonstrations show that ResViT outperforms competing methods based on CNNs and transformers in terms of qualitative observations and quantitative metrics.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Zhenda Xie, Zheng Zhang, Yue Cao, Yutong Lin, Jianmin Bao, Zhuliang Yao, Qi Dai, Han Hu
Summary: This paper presents SimMIM, a simple framework for masked image modeling. The study shows that the simple designs of each component have strong representation learning performance and achieve high accuracy on ImageNet-1K. Additionally, this approach addresses the data-hungry issue faced by large-scale model training.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2022)
Article
Geochemistry & Geophysics
Yule Duan, Hong Huang, Tao Wang
Summary: The article introduces a new method called GSMH, which is a geodesic-based sparse manifold hypergraph, to address the small sample problem in HSI data. This method utilizes geodesic distance and a geodesic-based neighborhood SR model to explore sparse correlations among different manifold neighborhoods, then constructs a pair of semisupervised hypergraphs to obtain nonlinear discriminative feature representation.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Xiaohan Ding, Xiangyu Zhang, Jungong Han, Guiguang Ding
Summary: This paper revisits large kernel design in modern CNNs and demonstrates that using a few large convolutional kernels can be more powerful. The authors propose RepLKNet, a CNN architecture with large kernel size, which closes the performance gap between CNNs and ViTs and shows scalability to big data and large models. The study also reveals the advantages of large-kernel CNNs in terms of effective receptive fields and shape bias.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2022)
Article
Geochemistry & Geophysics
Gong Cheng, Yanqing Yao, Shengyang Li, Ke Li, Xingxing Xie, Jiabao Wang, Xiwen Yao, Junwei Han
Summary: This article presents a two-stage oriented object detection method, called dual-aligned oriented detector (DODet), to address the spatial and feature misalignment problems. DODet uses an oriented proposal network to generate high-quality proposals and a localization-guided detection head to alleviate the feature misalignment between classification and localization. Extensive evaluations on multiple benchmarks show consistent and substantial improvements compared to baseline methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Rui Li, Shunyi Zheng, Chenxi Duan, Jianlin Su, Ce Zhang
Summary: The study introduces a linear attention mechanism (LAM) to address the issue of increasing memory and computational costs of the dot-product attention mechanism with large-scale inputs, enhancing the flexibility and versatility of integration between attention mechanisms and deep networks.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)