检索结果-内蒙古大学图书馆

Scale‐wise interaction fusion and knowledge distillation network for aerial scene recognition

CAAI Transactions on Intelligence Technology 2023年第4期8卷 1178-1190页

作者： Hailong Ning Tao Lei Mengyuan An Hao Sun Zhanxuan Hu Asoke K.Nandi School of Computer Science and Technology Xi'an University of Posts and TelecommunicationsShaanxi Key Laboratory of Network Data Analysis and Intelligent ProcessingXi'anChina Xi'an Key Laboratory of Big Data and Intelligent Computing Xi'anChina School of Electronic Information and Artificial Intelligence Shaanxi University of Science and TechnologyXi'anChina School of Computer Central China Normal UniversityWuhanChina Department of Electronic and Electrical Engineering Brunel University LondonLondonUK Xi'an Jiaotong University Xi'anChina

Aerial scene recognition(ASR)has attracted great attention due to its increasingly essential *** of the ASR methods adopt the multi‐scale architecture because both global and local features play great roles in ***,the existing multi‐scale methods neglect the effective interactions among different scales and various spatial locations when fusing global and local features,leading to a limited ability to deal with challenges of large‐scale variation and complex background in aerial scene *** addition,existing methods may suffer from poor generalisations due to millions of to‐belearnt parameters and inconsistent predictions between global and local *** tackle these problems,this study proposes a scale‐wise interaction fusion and knowledge distillation(SIF‐KD)network for learning robust and discriminative features with scaleinvariance and background‐independent *** main highlights of this study include two *** the one hand,a global‐local features collaborative learning scheme is devised for extracting scale‐invariance features so as to tackle the large‐scale variation problem in aerial scene ***,a plug‐and‐play multi‐scale context attention fusion module is proposed for collaboratively fusing the context information between global and local *** the other hand,a scale‐wise knowledge distillation scheme is proposed to produce more consistent predictions by distilling the predictive distribution between different scales during *** experimental results show the proposed SIF‐KD network achieves the best overall accuracy with 99.68%,98.74%and 95.47%on the UCM,AID and NWPU‐RESISC45 datasets,respectively,compared with state of the arts.

关键词： deep learning image analysis image classification information fusion

来源：评论

学校读者我要写书评

暂无评论

SAR Pulse Radio Frequency Interference Suppression Based on Modified Eigen Subspace Projection

SAR Pulse Radio Frequency Interference Suppression Based on ...

引用

2023 International Conference on Remote Sensing, Mapping, and Geographic Systems, RSMG 2023

作者： Liu, Mingyue Shu, Gaofeng Li, Ning School of Computer and Information Engineering Henan University Kaifeng475004 China Henan Key Laboratory of Big Data Analysis and Processing Kaifeng475004 China Henan Province Engineering Research Center of Spatial Information Processing Kaifeng475004 China

ISBN: (纸本)9781510668898

Synthetic aperture radar (SAR), as an active microwave remote sensor with high-resolution earth observation, is easily affected by radio frequency interference (RFI). Pulse radio frequency interference (PRFI), as a typical form of RFI, pollutes SAR raw data with bright stripes, which increases the difficulty of SAR image interpretation. However, the conventional eigen subspace projection (ESP) method decomposes the whole pulse into the interferences and the useful signals, resulting in the loss of useful signals in the non-interference positions with the pulse. To solve this problem, this paper proposes an improved SAR PRFI suppression method based on ESP. Firstly, the receiving pulses containing PRFI are detected and the specific position of PRFI in the pulse are determined by eigenvalue decomposition and energy accumulation in time-frequency domain, respectively. Secondly, the PRFI signal is separated from the useful signal by applying the ESP method only to the data where the PRFI is located. Finally, the PRFI signal projected by ESP is eliminated from the raw data. Simulation results shown that the proposed method can better protected useful signals and improve SAR image quality compared with conventional methods. © 2023 SPIE.

关键词： Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

Regional Winter Wheat Yield Estimation Based on CASA Model and Improved Harvest Index Using Remote Sensing images

Regional Winter Wheat Yield Estimation Based on CASA Model a...

引用

2023 International Conference on Remote Sensing, Mapping, and Geographic Systems, RSMG 2023

作者： Li, Qian Zhao, Jianhui Yang, Huijin Li, Ning School of Computer and Information Engineering Henan University Kaifeng475004 China Henan Province Engineering Research Center of Spatial Information Processing Kaifeng475004 China Henan Key Laboratory of Big Data Analysis and Processing Kaifeng475004 China

ISBN: (纸本)9781510668898

Accurately and timely grasping agricultural information at the regional scale helps to solve food security issues and formulate agricultural policies. Remote sensing images have the advantages of wide monitoring range and the ability to eliminate human interference. In recent years, they have been increasingly valued in crop yield estimation. The CASA model is used to estimate the net primary productivity (NPP) of crops and then combine it with the harvest index (HI) to estimate crop yields. However, most studies use a fixed HI for yield estimation, which can lead to low accuracy. In this context, the HI is studied and improved for winter wheat yield prediction in this research. The main contents are as follows: Firstly, time-series Sentinel-2 optical data and meteorological data were used as inputs of the CASA model to calculate the NPP of winter wheat. Then, HI was calculated using the fieldwork data and a method was proposed to improve it due to its geographic location-related characteristics. Finally, the calculated NPP and improved HI were used to estimate the winter wheat yield. The experimental results prove that the proposed method has higher accuracy than original methods, with determination coefficient(R2) of 0.595, root mean square error(RMSE) of 793.4 kg/ha, and mean absolute error(MAE) of 659.53 kg/ha. © 2023 SPIE.

关键词： Crops

来源：评论

学校读者我要写书评

暂无评论

Enhancing Surface Soil Moisture Estimation in Agricultural Fields: A Combined Approach of Improved WCM and CNN

Enhancing Surface Soil Moisture Estimation in Agricultural F...

引用

2023 International Conference on Remote Sensing, Mapping, and Geographic Systems, RSMG 2023

作者： Wang, Ran Zhao, Jianhui Li, Ning School of Computer and Information Engineering Henan University Kaifeng475004 China Henan Province Engineering Research Center of Spatial Information Processing Kaifeng475004 China Henan Key Laboratory of Big Data Analysis and Processing Kaifeng475004 China

ISBN: (纸本)9781510668898

Real-time and dynamic monitoring of soil moisture is critically essential for farming activities and crop yield estimation. This paper focuses on the study of surface soil moisture (SSM) in agricultural fields with different vegetation sparsity. Based on the improved water cloud model (WCM) and convolutional neural network (CNN), the SSM inversion of agricultural fields covered with winter wheat at different growth stages was achieved. Firstly, the vegetation cover factor was introduced into the WCM to separate the scattering contribution under crop cover from the direct scattering contribution from the bare ground. Then, the remote sensing data was used to derive numerous characteristic parameters to the SSM. Finally, the CNN model was established for the purpose of SSM inversion. The better correlation between the backscattering coefficients obtained by the improved WCM simulations and the measured SSM was shown. The coefficients of determination were 0.46 and 0.39 under VV and VH polarization, respectively. Better inversion accuracy can be obtained by combining the two parameters with other characteristic parameters as input data into the CNN model for SSM inversion. The coefficient of determination was 0.75, root mean square error was 2.51 vol.% and mean absolute error was 2.12 vol.%. Therefore, the method can effectively separate the effects of agricultural crops and bare ground on radar signals, which contributes a novel research approach and guidance for handling SSM inversion in agricultural fields covered by vegetation. © 2023 SPIE.

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

A Survey on Multimodal Dialogue Systems: Recent Advances and New Frontiers 5

A Survey on Multimodal Dialogue Systems: Recent Advances and...

引用

5th International Conference on Advanced Electronic Materials, Computers and Software Engineering, AEMCSE 2022

作者： Liu, Guangya Wang, Shiqi Yu, Jianxing Yin, Jian School of Artificial Intelligence Sun Yat-sen University Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou510006 China

ISBN: (数字)9781665484749

ISBN: (纸本)9781665484749

Recently, there has been growing interest in the field of multimodal dialogue systems. Different from traditional unimodal dialogue systems, our task needs to understand the context of multiple modalities before responding to users' utterances. In this paper, we present a detailed survey of the recent advances in multimodal dialogue systems and discuss some possible research directions. In particular, we categorize the dialogue systems into two basic tasks, including outputting textual responses and outputting visual responses. In these tasks, there are two main challenges, that is, the heterogeneity gap and the semantic gap. Then, we analyze the key techniques used to solve these challenges. Moreover, we review benchmark datasets and popular evaluation metrics comprehensively. Finally, we give some promising directions for future works. © 2022 IEEE.

关键词： Surveys

来源：评论

学校读者我要写书评

暂无评论

Detecting Ethereum Phishing Scams via Multi-transaction-view Graph Attention Network 29

Detecting Ethereum Phishing Scams via Multi-transaction-view...

引用

29th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2023

作者： Zhuo, Shaoxuan Li, Guang Wu, Weigang Bian, Jing Sun Yat-Sen University School of Computer Science and Engineering Guangzhou China Guangdong Province Key Laboratory of Big Data Analysis and Processing China Sun Yat-Sen University Guangdong Province Key Laboratory of Computational Science Guangzhou China

ISBN: (纸本)9798350330717

Cryptocurrency phishing scams is a significant treat to Ethereum, one of the most popular blockchain platforms. Most of existing Ethereum phishing detection methods are based on traditional machine learning or graph representation learning, which mostly rely on only statistical and structural features in local scope. In this paper, we propose Multi-transaction-view Graph Attention Network (MTvGAT), a novel phishing scam detection model that can make use of transaction patterns of different scopes. To obtain global-view information, we apply graph clustering and construct the global-view graph with multiple clusters, including all the nodes of the original transaction network. To obtain local view information, we apply neighborhood sampling, and construct local-view graphs with target nodes and their neighborhood nodes. Then, node features, edge features, and attention coefficients are aggregated to merge multi-view information into representation of nodes. We further combine global-view and local-view representations to finally identify phishing addresses from target nodes. Extensive experiments demonstrate that the proposed method can outperform existing ones with significant improvement. © 2023 IEEE.

关键词： Ethereum Graph neural network Phishing scams

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of Spaceborne SAR Radio Frequency Interference Detection and Suppression System

Design and Implementation of Spaceborne SAR Radio Frequency ...

引用

2023 International Conference on Remote Sensing, Mapping, and Geographic Systems, RSMG 2023

作者： Hu, Xingwang Zhang, Hengrui Li, Ning Guo, Zhengwei School of Computer and Information Engineering Henan University Kaifeng475004 China Henan Province Engineering Research Center of Spatial Information Processing Kaifeng475004 China Henan Key Laboratory of Big Data Analysis and Processing Kaifeng475004 China

ISBN: (纸本)9781510668898

Spaceborne synthetic aperture radar (SAR) is an active radar system carried on a satellite, with the help of synthetic aperture technology to achieve all-day, all-weather, high-resolution imaging, it has been widely used in military reconnaissance, resource exploration, deep space exploration and other fields. As the electromagnetic environment becomes more and more complex, spaceborne SAR is more likely to receive radio frequency interference (RFI) from other radar equipment in the working frequency band, which will affect the interpretation of SAR images and lead to image quality degradation. Although there are endless methods for RFI detection and suppression, each method has its limitations and can only deal with specific interference scenarios. A unified, robust and high-precision SAR interference detection and suppression platform is urgently needed. Based on SAR signal processing technology, this paper uses multiple methods to jointly design a unified high-robust system for SAR data containing different types of RFI, and realizes efficient detection and suppression of spaceborne SAR RFI. The popularization and application of this system is of great significance to improve the anti-interference ability of SAR system, improve the processing efficiency of a large number of SAR data in complex scenes and ensure the accuracy of data. © 2023 SPIE.

关键词： Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

PG-YOLO: A Lightweight Vehicle Detection Method for Drones 22

PG-YOLO: A Lightweight Vehicle Detection Method for Drones

引用

22nd IEEE International Symposium on Parallel and Distributed processing with Applications, ISPA 2024

作者： Liu, Yang Fan, Tengyu Mei, Kun Zhou, Liming Henan University Henan Key Laboratory of Big Data Analysis and Processing Henan Kaifeng475004 China Henan University School of Computer and Information Engineering Henan Kaifeng475004 China Huaihe Hospital of Henan University Department of Clinical Laboratory Henan Kaifeng475004 China

ISBN: (纸本)9798331509712

With the rapid development of drone technology, object detection technology is widely applied in the field of drone intelligent transportation. In response to the problems of low accuracy and high computational complexity in vehicle detection tasks on resource limited drones, this paper proposes a lightweight and efficient improved model PG-YOLO for drone vehicle detection based on the YOLOv7-tiny model. Firstly, replace the LeakyReLU activation function with the SiLU activation function to compensate for the lack of nonlinear expression in the LeakyReLU activation function, and improve the convergence speed and generalization of the model during training;Secondly, the feature extraction and fusion modules were optimized to enhance the network's perception and capture capabilities of the target, improve detection accuracy, and reduce the number of parameters;Then, lightweight convolution PConv and GSConv are introduced, and a lightweight module PPC based on PConv is proposed to improve the network's Neck layer, reducing the number of parameters and computation while maintain detection performance;Finally, the CBAM attention mechanism is introduced to improve the detection performance for the problem of dense vehicle targets. The improved network achieved an average detection accuracy of 60.1%, an increase of 12.1 percentage points compared to the original network, a reduction of 73.6% in parameter count, a reduction of 15.9% in computational complexity, and a detection speed of 154 FPS. Compared to the original network, the network is more lightweight and achieves higher detection accuracy, which can more effectively perform unmanned aerial vehicle detection tasks. © 2024 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

LLDB: Efficient Low-Light image Enhancement with Difffusion Bridge

LLDB: Efficient Low-Light Image Enhancement with Difffusion ...

引用

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Junlong Ma Conghan Yue Zhengwei Peng Dongyu Zhang Sun Yat-Sen University Guangzhou China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

This paper investigates a low-light image enhancement method based on the diffusion bridge framework. Currently, low-light image enhancement tasks still face challenges in noise reduction and detail restoration, and existing diffusion model methods are time-consuming and have unstable diffusion processes. We conducted an in-depth study of diffusion bridge theory, integrating the advantages of the end-to-end paradigm of diffusion bridge theory and employing a nonlinear activation network to further enhance the performance of low-light image enhancement tasks. Additionally, we used a Gamma correction module for fine-tuning low-light images, significantly improving performance with almost no extra computational cost. Experiments show that Low-Light image enhancement with Difffusion Bridge (LLDB) far surpasses other methods on LOLv1 and LOLv2 datasets. Code is available at https://***/M-Chase/LLDB

关键词： Bridges Computational modeling Noise reduction Diffusion processes Speech enhancement Signal processing Diffusion models image restoration image enhancement Faces

来源：评论

学校读者我要写书评

暂无评论

SymGaussian: Occluded Human Rendering with Multi-scale Symmetry Feature from Monocular Video

SymGaussian: Occluded Human Rendering with Multi-scale Symme...

引用

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Zekai Jiang Tong Duan Dongyu Zhang Sun Yat-Sen University Guangzhou China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

The growing demand for high-quality 3D human rendering in real-world applications highlights significant challenges. These challenges are particularly evident in dealing with occlusion in monocular video. Previous methods often rely on controlled datasets and overlook the inherent symmetry of the human body, leading to incomplete rendering in occluded areas. To address these limitations, we propose SymGaussian, a novel Gaussian Splatting-based approach for rendering occluded human from monocular video. We introduce Multi-scale Symmetry Feature to compensate for lost information in occluded areas, along with a Projective Texture Mapping method that efficiently encodes 2D appearance while preserving 3D perception. Experiments show that SymGaussian outperforms state-of-the-art methods in rendering quality, while achieving rapid training speed and real-time rendering capability exceeding 200 FPS.

关键词： Training Solid modeling Three-dimensional displays Streaming media Signal processing Rendering (computer graphics) Real-time systems Acoustics Speech processing

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：