Article
Engineering, Electrical & Electronic
Yimin Zhu, Kexin Yuan, Wenlong Zhong, Linlin Xu
Summary: The article presents an enhanced convolutional approach called SS-ConvNeXt for hyperspectral image classification. It introduces Spatial-ConvNeXt and Spectral-ConvNeXt blocks, as well as a spectral projection module, to better capture spatial and spectral information. Experimental results show that the proposed method achieves high classification accuracy while preserving class boundaries and reducing within-class noise.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Cuiping Shi, Haiyang Wu, Liguo Wang
Summary: This article proposes a positive feedback spatial-spectral correlation network based on spectral interclass slicing (PFSSC_SICS) to address the issues of lack of labeled samples and spectral similarity among classes in hyperspectral image classification (HSIC). The PFSSC_SICS method incorporates a spectral interclass slicing strategy to remove similar spectral signatures between classes and improve classification performance. It also introduces a positive feedback mechanism and a spatial-spectral correlation module to extract deeper and more features. Experimental results demonstrate that PFSSC_SICS outperforms state-of-the-art methods in HSIC.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Computer Science, Information Systems
Wenping Ma, Haoxiang Ma, Hao Zhu, Yating Li, Longwei Li, Licheng Jiao, Biao Hou
Summary: This paper proposes a spatial and spectral kernels generation network (SSKNet) to tackle the classification of hyperspectral images, achieving more efficient feature extraction and fusion through the generation of spatial and spectral convolution kernels. Experimental results demonstrate that this method outperforms existing ones in terms of classification accuracy and generalization performance.
INFORMATION SCIENCES
(2021)
Article
Geochemistry & Geophysics
Lei Li, Jihao Yin, Xiuping Jia, Sen Li, Bingnan Han
Summary: The letter proposes a data-driven joint spatial-spectral attention network (JSSAN) to extract more representative features from hyperspectral images by designing spatial-spectral attention blocks and inserting them into a CNN structure. The method outperforms several state-of-the-art algorithms in experimental results.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2021)
Article
Geochemistry & Geophysics
Ruoxi Song, Yining Feng, Wei Cheng, Zhenhua Mu, Xianghai Wang
Summary: This article proposes a novel bottleneck spatial-spectral transformer (BS2T) for capturing the long-range global dependencies of hyperspectral (HS) image pixels. It replaces convolutional operations with multihead spatial-spectral self-attention (MHS2A) to overcome the limitations of CNN-based HS image classification methods. A dual-branch HS image classification framework based on 3-D CNN and BS2T is defined for extracting local-global features of HS images. Experimental results demonstrate significant improvement compared to state-of-the-art methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Kai Yang, Hao Sun, Chunbo Zou, Xiaoqiang Lu
Summary: This paper introduces a cross-attention spectral-spatial network (CASSN) to address the rotation issue in hyperspectral image classification, by extracting spectral and spatial features to determine pixel categories.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Chemistry, Multidisciplinary
Yifan Si, Dawei Gong, Yang Guo, Xinhua Zhu, Qiangsheng Huang, Julian Evans, Sailing He, Yaoran Sun
Summary: In this paper, a segmentation framework based on the DeepLab v3+ neural network was proposed for hyperspectral imagery classification. The framework utilizes PCA for dimensionality reduction, combines spatial features extracted by DeepLab v3+ with spectral features, and employs an SVM classifier for fitting and classification. Experimental results demonstrate that the proposed framework outperforms traditional machine learning and deep-learning algorithms in hyperspectral imagery classification tasks.
APPLIED SCIENCES-BASEL
(2021)
Article
Environmental Sciences
Xin He, Yushi Chen, Zhouhan Lin
Summary: This study explores the application of Transformer-based models in hyperspectral image classification and proposes two new classification frameworks to address the issues of sequential data processing and spatial feature extraction. Experimental results show that the proposed models are competitive compared to state-of-the-art methods, indicating the new potential of Transformer in HSI classification.
Article
Geochemistry & Geophysics
Siyuan Hao, Yufeng Xia, Lijian Zhou, Yuanxin Ye, Wei Wang
Summary: In this study, a spectral and spatial feature fusion module (TransCNN) is proposed for hyperspectral image classification (HIC). By combining the advantages of CNN and Transformer, TransCNN can effectively extract spectral-spatial information and correlation of the three dimensions. Experimental results demonstrate that the proposed module achieves competitive performance on real hyperspectral images.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Computer Science, Information Systems
Tao Hu, Jing Yuan, Xiaodong Wang, Changxiang Yan, Xueping Ju
Summary: Marine oil spills pose a serious threat to the marine ecological environment. Hyperspectral remote sensing images (HRSI) technology has been proposed to accurately identify oil film on the marine surface. A spectral-spatial features extraction (SSFE) method based on convolutional neural networks (CNNs) was introduced to overcome the limitations of traditional spectrum analysis methods. Experimental results show that the proposed approach has achieved satisfactory performance in distinguishing oil spills.
Article
Engineering, Electrical & Electronic
Ningyang Li, Zhaohui Wang, Faouzi Alaya Cheikh, Mohib Ullah
Summary: This article proposes a novel spectral-similarity-based spatial attention module (S(3)AM) to emphasize relevant spatial areas in hyperspectral images. The proposed module measures spectral similarities using weighted Euclidean and cosine distances, and reweights the bands using full-band convolutional layers to alleviate the negative influence of spectral variability. Experimental results demonstrate the effectiveness of the module and the outstanding classification performance of the S(3)AM-Net model.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Cuiping Shi, Jingwei Sun, Liguo Wang
Summary: This paper proposes a spatial-spectral attention fusion network for hyperspectral image classification, which uses a four branch multiscale block and 3D-Softpool to extract and fuse spectral and spatial features. Experimental results show that the proposed method outperforms other classification methods when using a small number of training samples.
Article
Engineering, Electrical & Electronic
Huan Liu, Wei Li, Xiang-Gen Xia, Mengmeng Zhang, Chen-Zhong Gao, Ran Tao
Summary: This article addresses the issue of spectral shift in cross-scene hyperspectral imagery classification by proposing spectral shift mitigation (SSM) that includes amplitude shift mitigation (ASM) and adjacency effect mitigation (AEM). By reducing amplitude shift and spectral variation through amplitude normalization and weighted average spectral vector methods, and using a classifier trained with labeled samples from the source scene, superior classification performance is achieved on several cross-scene HSI data pairs.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Geochemistry & Geophysics
Uzair Aslam Bhatti, Zhaoyuan Yu, Jocelyn Chanussot, Zeeshan Zeeshan, Linwang Yuan, Wen Luo, Saqib Ali Nawaz, Mughair Aslam Bhatti, Qurat ul Ain, Anum Mehmood
Summary: This article introduces a novel HSI classification algorithm, LSPGF, which utilizes LSP and Gabor filtering to extract deeper features from original hyperspectral data, achieving higher classification accuracy compared to other algorithms.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Haoyang Yu, Zhen Xu, Ke Zheng, Danfeng Hong, Hao Yang, Meiping Song
Summary: In this paper, a multilevel spectral-spatial transformer network (MSTNet) is proposed for hyperspectral image classification. By introducing self-attentive and pure transformer encoders, the proposed method achieves efficient and accurate classification results.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Engineering, Multidisciplinary
Ke Xu, Wenzhuo Li, Chenghao Ji, Bing Liu
Summary: This paper introduces a tendon-driven robotic system composed of a robotic arm and an exoskeleton, and studies its control technology. Through the analysis of the kinematics and dynamics of the robot, and solving the problems of zero-force control and impedance control, the human-machine integration capability of the robot is improved, and the usage scenarios and application scope of the robot are expanded.
JOURNAL OF BIONIC ENGINEERING
(2023)
Review
Pharmacology & Pharmacy
Xianlin Qu, Bing Liu, Longgang Wang, Luguang Liu, Weizhu Zhao, Changlei Liu, Jishuang Ding, Siwei Zhao, Botao Xu, Hang Yu, Xiang Zhang, Jie Chai
Summary: In this study, a novel lncRNA called DACT3-AS1 was found to play a crucial role in gastric cancer (GC) by targeting the miR-181a-5p/SIRT1 axis to suppress cell proliferation, migration, and invasion. Additionally, DACT3-AS1 was transmitted from cancer-associated fibroblasts (CAFs) to GC cells via exosomes. Moreover, DACT3-AS1 enhanced the sensitivity of GC cells to oxaliplatin through SIRT1-mediated ferroptosis. Therefore, DACT3-AS1 may serve as a potential diagnostic and therapeutic target for GC.
DRUG RESISTANCE UPDATES
(2023)
Article
Geochemistry & Geophysics
Yifan Sun, Bing Liu, Xuchu Yu, Anzhu Yu, Pengqiang Zhang, Zhixiang Xue
Summary: For hyperspectral images (HSIs), the discrepancy of contiguous spectral information is crucial for ground object identification. The SpectralFormer, based on a transformer backbone, effectively captures the long-term dependence of the spectrum and enhances the performance of spectral feature methods.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Chemistry, Multidisciplinary
Danyang Hong, Chunping Qiu, Anzhu Yu, Yujun Quan, Bing Liu, Xin Chen
Summary: Building extraction and change detection from remote sensing imagery are important and challenging tasks with application potential. We propose a multi-task learning network using Swin transformer as a shared backbone network for simultaneous building extraction and change detection. Experimental results on the WHU-CD dataset demonstrate the effectiveness of the proposed method.
APPLIED SCIENCES-BASEL
(2023)
Article
Engineering, Chemical
Bing Liu, Kun Dong, Ru-Jia Chen, Song-Tao Chang, Guang-Wen Chu, Liang-Liang Zhang, Hai-Kui Zou, Bao-Chang Sun
Summary: This study investigates the solid-liquid mass-transfer performance of suspension in a rotating packed bed (RPB) and develops a calculation model for the mass-transfer coefficient. The effects of various operating conditions on the RPB are examined, including dispersion time, rotating speed, mesh packing thickness, liquid volume flow rate, and solid loading. The results show that the increase in rotating speed and mesh packing thickness leads to an increase in the mass-transfer coefficient. However, the solid loading has little influence. A nondimensional equation for the Sherwood number is established and the predicted values agree well with experimental values.
INDUSTRIAL & ENGINEERING CHEMISTRY RESEARCH
(2023)
Article
Remote Sensing
Bing Liu, Xibing Zuo, Anzhu Yu, Yifan Sun, Ruirui Wang
Summary: This study proposes a semi-supervised deep-learning method based on multi-view consistency to improve the classification accuracy of hyperspectral images using a few labelled samples. The method builds a classifier based on a residual network and introduces an attention mechanism to enhance classification performance. The unsupervised loss function is used to train the model and fully utilize unlabelled samples for improved accuracy.
REMOTE SENSING LETTERS
(2023)
Article
Remote Sensing
Bing Liu, Anzhu Yu, Xibing Zuo, Ruirui Wang, Chunping Qiu, Xuchu Yu
Summary: Deep learning with convolutional neural networks has been successful in high-resolution remote-sensing image change detection. However, these networks cannot capture long-range dependencies in a scene. To address this, we propose a hierarchical transformer model that can better describe long-range dependencies and improve the precision of change detection in high-resolution remote sensing images.
EUROPEAN JOURNAL OF REMOTE SENSING
(2023)
Article
Environmental Sciences
Wenkai Liu, Bing Liu, Peipei He, Qingfeng Hu, Kuiliang Gao, Hui Li
Summary: A novel small sample classification method based on rotation-invariant uniform local binary pattern (RULBP) features and a graph-based masked autoencoder is proposed to deal with the problem of small samples in hyperspectral image classification. The RULBP features of hyperspectral images are extracted and a graph is constructed using the k-nearest neighbor method. Self-supervised learning is conducted on the graph to extract features suitable for small sample classification. After training, a small number of samples are used to fine-tune the graph convolutional network to complete the classification.
Article
Environmental Sciences
Kuiliang Gao, Anzhu Yu, Xiong You, Chunping Qiu, Bing Liu, Wenyue Guo
Summary: This paper proposes a general-purpose representation learning method for cross-domain HSI classification, which utilizes a three-level distillation strategy to transfer knowledge from multiple models and quickly adapt to different target domains with small samples.
Article
Computer Science, Information Systems
Wenjun Huang, Qun Sun, Anzhu Yu, Wenyue Guo, Qing Xu, Bowei Wen, Li Xu
Summary: This study investigates the recognition of point symbols on scanned topographic maps using a deep learning-based method. The proposed deep convolutional neural network (DCNN) model, along with atrous spatial pyramid pooling (ASPP) and data augmentation methods, improves the accuracy and efficiency of recognition compared to classical algorithms. This research contributes to algorithm development and the evaluation of geographic elements extracted from topographic maps.
ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION
(2023)
Article
Environmental Sciences
Kuiliang Gao, Anzhu Yu, Xiong You, Chunping Qiu, Bing Liu, Fubing Zhang
Summary: This paper focuses on the problem of semi-supervised domain adaptation (SSDA) segmentation of remote sensing images (RSIs). A novel cross-domain multi-prototype constraint and contradictory structure learning mechanism are proposed, and self-supervised learning is adopted to increase the number of target samples involved. Experimental results demonstrate the effectiveness of the proposed method in improving segmentation performance and narrowing the gap with supervised methods.
Review
Environmental Sciences
Anzhu Yu, Yujun Quan, Ru Yu, Wenyue Guo, Xin Wang, Danyang Hong, Haodi Zhang, Junming Chen, Qingfeng Hu, Peipei He
Summary: This paper summarizes and reviews the methods and challenges of supervised learning with a small number of labeled samples in semantic segmentation tasks of remote sensing images under deep learning framework. It also involves different training methods such as self-supervised learning, semi-supervised learning, weakly supervised learning, and few-shot methods, as well as the solutions to cross-domain challenges.