检索结果-内蒙古大学图书馆

Lightweight image super-resolution via an adaptive information fusion attention network

Soft Computing 2025年 1-15页

作者： He, Yi Huan, Hai Zou, Nan Zhang, Yi Xie, Yaqin Wang, Chao School of Artificial Intelligence Nanjing University of Information Science & Technology Nanjing210044 China School of Electronic & Information Engineering Nanjing University of Information Science & Technology Nanjing210044 China Jiangsu Key Laboratory of Meteorological Observation and Information Processing Nanjing University of Information Science and Technology Nanjing210044 China Nanjing Nanji Agricultural Machinery Science and Technology Research Institute Nanjing210019 China

The Convolutional Neural Networks (CNNs) have achieved outstanding performance in the field of Image Super-Resolution (SR). However, many existing methods focus excessively on increasing network depth, which leads to an excessive number of parameters and computational complexity. Some methods, aimed at reducing model capacity, significantly reduce the receptive field of the network, and employ a single upsampling method, resulting in suboptimal model performance. To better balance model size and performance, this paper proposes a lightweight Adaptive information Fusion Attention Network (AIFAN). Specifically, a Dual Path Block (DPB) is designed to adaptively aggregate multi-scale features from different paths through the Adaptive Path Fusion Block (APFB). To take full advantage of the hierarchical features, all DPBs are connected using a dense connection mode. Meanwhile, a receptive field attention (RFA) module is proposed to increase the receptive field of the network and force the network to pay attention to the region with more information in the feature. In addition, a hybrid up-sampling module is designed. The Adaptive Multi-scale Up-sampling (AMSU) module and Feature Update Up-sampling Block (FUUB) are combined in the module to obtain better reconstruction results. The experimental results show that the proposed AIFAN achieves an effective balance between performance and model complexity. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： information fusion

来源：评论

学校读者我要写书评

暂无评论

A fine segmentation model of flue-cured tobacco’s main veins based on multi-level-scale features of hybrid fusion

引用

Soft Computing 2024年第17-18期28卷 10537-10555页

作者： Xu, Biao Liu, Xiaobao Gu, Wenjuan Liu, Jia Wang, Hongcheng Faculty of Mechanical and Electrical Engineering Kunming University of Science and Technology Kunming650500 China School of Electronic Information Engineering China West Normal University Nanchong637000 China

Flue-cured tobacco (FCT) can be classified into upper (B), middle (C), and lower (X) parts based on characteristics such as the FCT's main veins, leaf shape, color, and thickness. Accurately measuring the geometric parameters of the main veins is crucial for identifying the different parts. However, this task has proven to be challenging. Therefore, segmenting the main veins is a prerequisite to reducing calculation errors and improving the precision of part identification. To obtain enough semantic information and improve segmentation accuracy, we propose a fine segmentation model (MSHF-Net) of FCT's main veins based on multi-level-scale features of hybrid fusion. Firstly, MobileNetV2 with a dilated convolution layer (DMobileNetV2) is selected as the backbone network for feature extraction, which optimizes training and inference speed to minimize computing costs. Subsequently, Hybrid Fusion Atrous Spatial Pyramid Pooling (HFASPP) is designed to be the strengthened backbone module for capturing more high-level semantic information, effectively preventing intermittent segmentation of some main veins. Additionally, considering the low proportion of main vein targets in the original image, the double shallow feature branches (DSFBS) are included to obtain more low-level semantic information. Finally, a channel attention mechanism (ECANet) is added to enhance useful information and eliminate redundant information after the hybrid fusion of high-low-level semantic information, preventing mis-segmentation of regions. Experimental validation demonstrates the efficiency of the MSHF-Net, with parameters of only 7.92 M, thus ensuring minimal computational requirements. The model achieves an impressive mean intersection over union (MIoU) of 85.57% and mean pixel accuracy (mPA) of 93.10% on a diverse test set of FCT parts. When applied to segment main veins in a 2296 × 1548 × 3 tobacco image, the model takes just over 0.1 s. It is noteworthy that none of the 291 randomly segmen

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

A Noise Interaction Kalman Filter Based on Variational Bayesian for Handling Non-Gaussian Measurement and Process Noise

A Noise Interaction Kalman Filter Based on Variational Bayes...

引用

2024 IEEE International Conference on Unmanned Systems, ICUS 2024

作者： Chen, Jian Peng, Bei He, Jiacheng Wang, Gang School of Mechanical and Electrical Engineering University of Electronic Science and Technology of China Chengdu China School of Information and Communication Engineering University of Electronic Science and Technology of China Chengdu China

ISBN: (纸本)9798350384185

Underwater target tracking for unmanned systems has always played a crucial role. However, target measurements and process transitions are contaminated by non-Gaussian noise due to the underwater environment and the intrinsic system of the target, thereby degrading tracking performance. This paper investigates a filter capable of simultaneously mitigating measurement noise and unknown process noise. Specifically, underwater measurement noise is modeled as a Gaussian Mixture Model (GMM) with multiple noise components. Different component parameters form models to handle observational noise, and each model is assigned a different probability. Variational Bayesian (VB) inference is employed within each model to approximate the parameters of the unknown process noise distribution, and finally, fusion is performed based on noise probabilities. Simulations conducted with an Unmanned Underwater Vehicle (UUV) demonstrate that the proposed algorithm outperforms existing major algorithms when both measurement and process noises follow non-Gaussian distributions. © 2024 IEEE.

关键词： Kalman filters

来源：评论

学校读者我要写书评

暂无评论

A 12-Unit Asymmetric Mirror-Coupled Loop Antenna for 5G Smartphones

引用

Progress In Electromagnetics Research C 2024年 145卷 141-152页

作者： Ren, Wanying Wang, Zhonggen Nie, Wenyan Mu, Weidong Li, Chenlu Wang, Mingqing You, Wenshi School of Electrical and Information Engineering Anhui University of Science and Technology Huainan232001 China School of Mechanical and Electrical Engineering Huainan Normal University Huainan232001 China School of Electronic and Information Engineering Nanjing University of Aeronautics and Astronautics Nanjing211106 China School Electrical and Information Engineering Hefei Normal University Hefei230061 China

This paper introduces a 12-element asymmetric mirror-coupled loop antenna for integration into 5G smartphones. The proposed antenna includes six identical asymmetrically mirrored (AM)-coupled building blocks, each consisting of two gap-coupled loop antennas, with each building block dimensioning a mere 12 × 7 mm2. The unique architecture of the proposed antenna yields an isolation performance exceeding 14 dB without the necessity for ancillary decoupling elements and effectively covers the dual-band 5G mobile frequency bands of 3.3–3.6 GHz and 4.8–5.0 GHz. The antenna was optimized using simulation software HFSS, and the results indicate an antenna’s envelope correlation coefficient of less than 0.016 and an efficiency range of 72%–81%. Finally, the performance of single-hand and double-hand handset smartphone modes is discussed, which still exhibit good radiation and MIMO performance under both modes, demonstrating their stability in practical applications. Simulation and measurement results indicate that the proposed 12-element MIMO antenna holds great promise for 5G smartphone applications. © 2024, Electromagnetics Academy. All rights reserved.

关键词： Loop antennas

来源：评论

学校读者我要写书评

暂无评论

Sketch-based 3D Model Retrieval with Cross-Modal Representation 6

Sketch-based 3D Model Retrieval with Cross-Modal Representat...

引用

6th ACM International Conference on Multimedia in Asia, MMAsia 2024

作者： Yang, Hairui Wang, Ning Wang, Zhihui Wang, Lei University of Electronic Science and Technology of China Zhejiang Huzhou China International School of Information Science and Engineering Dalian University of Technology Liaoning Dalian China Software School Dalian University of Technology Liaoning Dalian China

ISBN: (纸本)9798400712739

Sketch-based 3D model retrieval is a challenging task in computer vision, requiring simultaneous solutions to discriminative feature learning and cross-modal semantic *** methods typically project sketches and 3D models into a shared feature space for both feature learning and semantic alignment, often resulting in the sub-optimal match of query sketch and retrieved 3D *** paper proposes a novel sketch-based 3D model retrieval method based on cross-modal ***, discriminative features of 3D models are learned without interference from ***, sketch feature learning and semantic alignment are guided by learned prototypes of 3D model ***, cross-modal representations generated by vision-language pre-trained models are introduced as category similarity priors to enhance the discriminative nature of sketch *** results demonstrate that the proposed method outperforms existing methods in retrieval performance, achieving a 2.5% improvement in mean average precision (mAP) on SHREC 2016. © 2024 Copyright held by the owner/author(s).

关键词： 3D modeling

来源：评论

学校读者我要写书评

暂无评论

CMCF-Net: An End-to-End Context Multiscale Cross-Fusion Network for Robust Copy-Move Forgery Detection

引用

IEEE Transactions on Multimedia 2024年 26卷 6090-6101页

作者： Xiong, Lizhi Xu, Jianhua Yang, Ching-Nung Zhang, Xinpeng Nanjing University of Information Science and Technology School of Computer Science Nanjing210044 China Jiangsu Collaborative Innovation Center on Atmospheric Environment and Equipment Technology Engineering Research Center of Digital Forensics Ministry of Education Nanjing210044 China Nanjing University of Information Science and Technology Software Engineering Nanjing210044 China National Dong Hwa University Department of Computer Science and Information Engineering Hualien974301 Taiwan Fudan University School of Computer Science Shanghai200433 China

Image copy-move forgery detection (CMFD) has become a challenging problem due to increasingly powerful editing software that makes forged images increasingly realistic. Existing algorithms that directly connect multiple scales of features in the encoder part may not effectively aggregate contextual information, resulting in poor performance. In this paper, an end-to-end context multiscale cross-fusion network (CMCF-Net) is proposed to detect image copy-move forgery. The proposed network consists of a multiscale feature extraction fusion (MSF) module and a multi-information fusion decoding (MFD) module. Multiscale information is efficiently extracted and fused in the MSF module utilizing stacked-scale feature fusion, which improves the network's forgery localization ability on objects of different scales. The MFD module employs contextual information combination and weighted fusion of multiscale information to guide the network in obtaining relevant clues from correlated information at multiple different scales. Experimental results and analysis have demonstrated that the proposed CMCF-Net achieves the best localization results with higher robustness. © 1999-2012 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

A Swarm Formation Control System for UAV-Enabled Internet of Things with Hybrid Path Planning

引用

IEEE Internet of Things Journal 2024年第12期12卷 18484-18494页

作者： Zhang, Xiaoying Zhang, Jiawei Mu, Zhihao Liu, He Xiao, Lei Wang, Xiaojie Ning, Zhaolong National University of Defense Technology College of Electronic Science and Technology Changsha410003 China Chongqing University of Posts and Telecommunications School of Communications and Information Engineering Chongqing400065 China

To address the problem of real-time path planning for UAV (Unmanned Aerial Vehicle) formations in complex environments, this paper proposes a hybrid path planning control system tailored for UAV-enabled Internet of Things (IoT) utilizing NFC (Near-Field Communication) among them. We model the UAV formation problem as an undirected graph, and first design a virtual leader-follower method based on the consistency principle of distributed systems. After that, we present an innovative hybrid path planning method (named ASAP-A*), combining A∗ and artificial potential field methods. It effectively reduces redundant points in the path and optimizes the UAV trajectory by B-spline smoothing to meet UAV trajectory requirements. Finally, we construct a user-friendly UAV swarms formation control system based the proposed methods, tackling the complexity of operating existing systems, and verifying the effectiveness of our solutions. © 2014 IEEE.

关键词： Motion planning

来源：评论

学校读者我要写书评

暂无评论

Generalized Minimum Error Entropy Unscented Kalman Filter based on Variational Bayesian

Generalized Minimum Error Entropy Unscented Kalman Filter ba...

引用

2024 IEEE International Conference on Unmanned Systems, ICUS 2024

作者： He, Jiacheng Feng, Zhenyu Wang, Gang Peng, Bei School of Mechanical and Electrical Engineering University of Electronic Science and Technology of China Chengdu China School of Information and Communication Engineering University of Electronic Science and Technology of China Chengdu China

ISBN: (纸本)9798350384185

In this paper, a new adaptive outlier-robust filter is developed to address the nonlinear filtering problem with time-varying and unknown heavy-tailed measurement noises. The generalized minimum entropy criterion is employed to offer greater flexibility in noise adaptation, and the variational Bayesian (VB) technique is utilized to further infer the noise statistics online. The fixed-point iteration strategy is employed to address the coupling between the variational inference and the entropy criterion optimization. A target tracking example is conducted to verify the superiority of the developed algorithm. The results show that the robust nonlinear filtering algorithm developed using the GMEE criterion has strong noise suppression capability. © 2024 IEEE.

关键词： Wiener filtering

来源：评论

学校读者我要写书评

暂无评论

SCAF-DG: A Multi-site Medical Image Denoising with A Domain-Generalized Spatial-Channel Attention Fusion 6

SCAF-DG: A Multi-site Medical Image Denoising with A Domain-...

引用

6th IEEE International Conference on Pattern Recognition and Artificial Intelligence, PRAI 2023

作者： Fiasam, Linda Delali Rao, Yunbo Sey, Collins Klugah-Brown, Benjamin Tettey, Obed Nartey Aggrey, Esther Stacy E.B. Agyemang, Isaac Osei Adjei-Mensah, Isaac Yang, Yang University of Electronic Science Technology of China School of Information and Software Engineering Chengdu China University of Electronic Science Technology of China School of Life Science and Technology Chengdu China University of Electronic Science Technology of China School of Information and Communication Engineering Chengdu China Sichuan Institute of Computer Sciences Chengdu China

ISBN: (纸本)9798350325485

Images from multiple medical sites usually contain varying noise levels that can affect the generalization performance of the denoising models. Domain generalization (DG), which seeks to learn a model that can generalize to an unseen test domain, has not been extensively explored in medical image denoising for multi-site images. We present an approach for multi-site medical image denoising, which addresses the issue of varying noise levels across different imaging sites. Our proposed method, called SCAF-DG (Spatial-Channel Attention Fusion for Domain Generalization), employs two attention units to extract both domain-specific channel features and mutually-invariant spatial features from the noisy images. These features are then combined to form a rich feature space. Additionally, we utilized skip connections to preserve structural information during the denoising process. We evaluate SCAF-DG on publicly available brain magnetic resonance image (MRI) datasets from three different imaging sites, and demonstrate an average increase of 1.01 in peak signal-To-noise ratio (PSNR) and 0.016 in structural similarity index measure (SSIM) for denoising images. Our results demonstrate that SCAF-DG has the potential to generalize well to new, unseen test domains, and outperforms existing state-of-The-Art denoising models on multi-site medical images. © 2023 IEEE.

关键词： Image denoising

来源：评论

学校读者我要写书评

暂无评论

Determining geocenter motion using combined ground and spaceborne GPS observations with ambiguity resolution

引用

Advances in Space Research 2025年第11期75卷 7903-7916页

作者： Guo, Shiwei Fan, Lei Li, Zongnan Fang, Xinqi Huo, Chenshu Shi, Chuang School of Space and Earth Sciences Beihang University Beijing102206 China Key Laboratory of Navigation and Communication Fusion Technology Ministry of Industry and Information Technology Beijing100191 China College of Electronic Science National University of Defense Technology Changsha410005 China School of Electronic and Information Engineering Beihang University Beijing100191 China

Owing to the dense global tracking network and the abundant satellites providing continuous observations, the Global Navigation Satellite System (GNSS) has the potential to measure geocenter motion. However, the high altitude of GNSS satellites and the estimation of phase ambiguities compromise the quality of the estimated geocenter coordinates (GCC). This study conducted a combined zero-difference processing of ground observations from 98 global stations and spaceborne observations from four low Earth orbiters (LEOs), investigating the impact of integer ambiguity resolution and LEO configuration on GPS-derived GCC estimates. The ambiguity resolution in the double-difference mode between ground stations and one LEO can significantly improve the observability of GCC estimates, reducing the formal errors by 69.4 %, 69.4 % and 55.0 % for the X, Y and Z components, respectively. The faster relative motion between LEO and GPS satellites contributes to a faster change of tracking geometry. Integrating four LEOs into ground network, the formal errors of GCC parameters are reduced by 77.8 %, 78.2 % and 50.5 % for the X, Y and Z components, respectively. For the most concerned GCC Z component, its correlation with the B1C empirical parameter is reduced from 0.69 to 0.41, demonstrating that the GCC Z component is more separable from the orbital parameters. Besides, the 5th and 7th draconitic harmonics of the Z component are mitigated by 67.0 % and 73.5 %, respectively. The derived annual signal shows a good consistency with the external SLR-based solution, and the amplitude differences are only 0.9 mm, 1.6 mm and 1.2 mm for the X, Y and Z components, respectively. © 2025

关键词： Global positioning system

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：