Related references
Note: Only part of the references are listed.
Article
Engineering, Electrical & Electronic
Wanjie Lu et al.
Summary: This study proposes a convolution neural network transformer hybrid model for efficient object detection in UAV images. The model has three advantages that contribute to improving object detection performance. Firstly, a cross-shaped window transformer is used to obtain image features at different levels, enabling multiscale object detection. Secondly, a hybrid patch embedding module is constructed to extract and utilize low-level information such as edges and corners. Finally, a slicing-based inference method is used to fuse the inference results of the original image and sliced images, improving small object detection accuracy without modifying the original network. Experimental results demonstrate that the proposed method outperforms popular and state-of-the-art object detection methods.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2023)
Article
Engineering, Electrical & Electronic
Ang Li et al.
Summary: This paper proposes a cross-modal knowledge distillation (CKD) enabled object detection paradigm for UAV-based target detection. It achieves comparable detection performance with multi-modal techniques while requiring less computational resources.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2023)
Article
Geochemistry & Geophysics
Jiaqing Zhang et al.
Summary: In this article, the authors propose SuperYOLO, an accurate and fast object detection method for remote sensing images. By fusing multimodal data and utilizing assisted super resolution learning, SuperYOLO achieves high-resolution object detection on multiscale objects while considering the computation cost. Experimental results show that SuperYOLO outperforms state-of-the-art models in terms of accuracy and computational efficiency.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Yue Cao et al.
Summary: Multimodal object detection has been a popular research topic in recent years. In this paper, a novel lightweight fusion module called CSSA is proposed to efficiently fuse inputs from different modalities. Experimental results demonstrate its excellent performance in improving detection accuracy.
2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW
(2023)
Article
Geochemistry & Geophysics
Xiumei Chen et al.
Summary: This letter proposes a local-global mutual learning (LML) approach to capture both the global and local features of remote sensing scene classification (RSSC). The method generates local regions by highlighting semantic areas in the original image and uses a two-branch architecture to extract features for the local regions and global image. Experimental results demonstrate the effectiveness of the proposed method.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Engineering, Electrical & Electronic
Jung Uk Kim et al.
Summary: This paper proposes a new uncertainty-aware multispectral pedestrian detection framework to address the issues of miscalibration and modality discrepancy. Experimental results show that the proposed method outperforms existing state-of-the-art methods.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Environmental Sciences
Qingwang Wang et al.
Summary: In the field of remote sensing image applications, object detection using RGB and infrared images is an important technology. This study proposes a redundant information suppression network (RISNet) to enhance the fusion of complementary information between RGB and infrared images. Experimental results show that the proposed method outperforms state-of-the-art approaches, especially in challenging conditions.
Article
Environmental Sciences
Chujie Xu et al.
Summary: This paper explores a cross-domain ship detection task, adapting the detector from labeled optical images to unlabeled SAR images. A multi-level alignment network is proposed to achieve cross-domain detection and reduce domain shift. Experimental results demonstrate the effectiveness of the method.
Article
Engineering, Electrical & Electronic
Yiming Sun et al.
Summary: Researchers have constructed a large-scale drone-based RGB-infrared vehicle detection dataset and proposed an uncertainty-aware cross-modality vehicle detection framework to improve detection performance in challenging conditions. The experiments show that the proposed method performs well in complex environments.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Jian Ding et al.
Summary: This paper presents a large-scale DOTA dataset for object detection in aerial images, along with comprehensive baselines and a code library. The dataset and evaluations provided can facilitate the design of robust algorithms and reproducible research in the field of object detection in aerial images.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Computer Science, Artificial Intelligence
Fang Qingyun et al.
Summary: This study proposes a novel multispectral feature fusion approach, which improves the perception ability of detection algorithms by cross-modality fusing complementary information. It achieves robust and reliable performance in applications such as nighttime detection.
PATTERN RECOGNITION
(2022)
Proceedings Paper
Automation & Control Systems
Xiaoxiao Yang et al.
Summary: This paper proposes an effective and efficient cross-modality fusion module called BAA-Gate for multispectral pedestrian detection. The module extracts informative features and recalibrates representations based on the attention mechanism, and optimizes the features of two modalities through a bi-direction multi-stage fusion strategy. Experimental results on the KAIST dataset demonstrate the superior performance and satisfactory speed of the proposed method.
2022 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2022)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Wentong Li et al.
Summary: This paper proposes an adaptive points learning approach for aerial object detection, which captures the geometric information of arbitrary-oriented instances. Experimental results demonstrate the efficacy of the proposed method.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022)
(2022)
Article
Computer Science, Artificial Intelligence
Xiangtao Zheng et al.
Summary: Visible-infrared person re-identification is a challenging task due to the significant differences between images captured in different spectra. This paper proposes a partially interactive collaboration method to reduce the modality gap, achieving impressive results through the collaborative shallow layers and shared deep layers architecture.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Article
Engineering, Electrical & Electronic
Wei Liu et al.
Summary: This paper proposes a tassel detection algorithm based on UAV imagery, which achieves better performance in small-size tassel detection. The algorithm adopts novel techniques such as the bidirectional feature pyramid network and the robust attention module.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Jingqian Xue et al.
Summary: This article proposes a Dual network structure with InterweAved Global-local feature hierarchy based on the TRansformer architecture (DIAG-TR) for object detection in remote sensing images. It addresses the limitations of Transformer-based object detection in modeling large scale variation and difficult training. Experimental results demonstrate that DIAG-TR outperforms the original method in terms of mean average precision and convergence time, and surpasses state-of-the-art methods, showcasing its great potential in the field of earth observation.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Computer Science, Artificial Intelligence
Xiangtao Zheng et al.
Summary: This paper proposes a rotation-invariant attention network (RIAN) for HSI classification, which extracts rotation-invariant spectral-spatial features using center spectral attention and rectified spatial attention modules. Experimental results show that RIAN performs well on HSIs with spatial rotation.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Article
Computer Science, Artificial Intelligence
Zhaowei Cai et al.
Summary: In object detection, the commonly used IoU threshold of 0.5 can lead to noisy detections, and performance may degrade for larger thresholds. The Cascade R-CNN architecture addresses this issue by training detectors sequentially with increasing IoU thresholds and eliminating quality mismatches at inference, resulting in state-of-the-art performance and significant improvement in high-quality detection across various datasets. The model is also generalized to instance segmentation, achieving nontrivial improvements over existing methods.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Xinyu Jia et al.
Summary: This study introduces a visible-infrared paired dataset LLVIP for low-light vision tasks. Experimental results demonstrate the complementary effect of fusion on image information and reveal the deficiencies of existing algorithms in very low-light conditions.
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS (ICCVW 2021)
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Heng Zhang et al.
Summary: The proposed novel attentive multispectral feature fusion approach utilizes deep learning architecture to dynamically weigh and fuse multispectral features guided by inter- and intra-modality attention modules, significantly improving pedestrian detection accuracy at a low computation cost based on experiments on two public datasets.
2021 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2021)
(2021)
Article
Computer Science, Artificial Intelligence
Hao Sun et al.
Summary: In this study, an end-to-end fully convolutional segmentation network (FCSN) is proposed to simultaneously identify land-cover labels of all pixels in a HSI cube. The study also introduces a fine label style to label all pixels of HSI cubes for detailed spatial land-cover distributions and a HSI cube generation method to improve the diversity of spatial land-cover distributions. Experimental results demonstrate that FCSN has superior generalization capabilities to the changes of spatial land-cover distributions.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)
Article
Geochemistry & Geophysics
Yuanlin Zhang et al.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2020)
Article
Computer Science, Artificial Intelligence
Lu Zhang et al.
INFORMATION FUSION
(2019)
Article
Computer Science, Artificial Intelligence
Dayan Guan et al.
INFORMATION FUSION
(2019)
Article
Computer Science, Artificial Intelligence
Chengyang Li et al.
PATTERN RECOGNITION
(2019)
Proceedings Paper
Computer Science, Artificial Intelligence
Jing Nie et al.
2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019)
(2019)
Article
Computer Science, Artificial Intelligence
Shaoqing Ren et al.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2017)
Article
Computer Science, Information Systems
Sebastien Razakarivony et al.
JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION
(2016)
Article
Computer Science, Artificial Intelligence
Olga Russakovsky et al.
INTERNATIONAL JOURNAL OF COMPUTER VISION
(2015)
Article
Computer Science, Artificial Intelligence
Piotr Dollar et al.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2012)