检索结果-内蒙古大学图书馆

2020 International Conference on Geomatics in the Big Data Era, ICGBD 2020

作者： He, Y.R. Ma, W.W. Wang, X.R. Dai, J.Q. Zheng, J.L. College of Computer and Information Engineering Xiamen University of Technology Xiamen China Key Laboratory of Ecological Environment Information Atlas Fujian Provincial University Putian University Putian China

The power patrol has been completed by manual field investigation, which is inefficient, costly and unsafe. In order to extract the height of the power line and its surrounding ground objects more quickly and conveniently, and better service for power line patrol. This paper uses remote sensing data of unmanned aerial vehicle to carry out aerial triangulation, stereo model establishment and binocular stereo vision height extraction base on MapMatrix software, then obtains the power line height analysis chart. Then LiDAR point cloud data is used to verify the accuracy of the power line height analysis chart. The results show that this method not only meets the standard of power line patrol, but also improves the efficiency and quality of power line patrol. © 2020 Y. R. He et al.

关键词： Stereo image processing

来源：评论

学校读者我要写书评

暂无评论

A Fraud Detection Approach Based on Combined Feature Weighting 10th

A Fraud Detection Approach Based on Combined Feature Weighti...

引用

10th EAI International Conference on Cloud Computing, CloudComp 2020

作者： Liu, Xiaoqian Yu, Chenfei Xia, Bin Gu, Haiyan Wang, Zhenli Department of Computer Information and Cyber Security Jiangsu Police Institute Nanjing210031 China Jiangsu Key Laboratory of Big Data Security and Intelligent Processing Nanjing University of Posts and Telecommunications Nanjing210023 China Jiangsu Electronic Data Forensics and Analysis Engineering Research Center Nanjing China Jiangsu Provincial Public Security Department Key Laboratory of Digital Forensics Nanjing China

ISBN: (纸本)9783030699918

Data mining technology has yielded fruitful results in the area of crime discovery and intelligent decision making. Credit card is one of the most popular payment methods, providing great convenience and efficiency. However, due to the vulnerabilities of credit card transactions, criminals are able to commit fraud to infringe on the interests of the state and citizens. How to discover potential fraudsters while guaranteeing high efficiency becomes an extremely valuable problem to solve. In this work, we talk about the advantages and disadvantages of different models to detect credit card fraud. We first introduce the data preprocessing measures for handling imbalanced fraud detection dataset. Then we compare related models to implement fraudster recognition. We also propose a feature selection approach based on combined feature weights. Some future research interests are also envisioned. © 2021, ICST Institute for computer Sciences, Social Informatics and Telecommunications Engineering.

关键词： Crime

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Time Series Anomaly Detection under Data Contamination

Unsupervised Time Series Anomaly Detection under Data Contam...

引用

IEEE International Symposium on information (IT) in Medicine and Education, ITME

作者： Xiaohui Lin Zuoyong Li Xunhua Huang Xinwei Chen Haoyi Fan College of Computer and Data Science Fuzhou University Fuzhou China Fujian Provincial Key Laboratory of Information Processing and Intelligent Control College of Computer and Control Engineering Minjiang University Fuzhou China School of Computer Science and Technology Harbin University of Science and Technology Harbin China Fujian Engineering and Research Center of New Chinese Lacquer Materials Minjiang University Fuzhou China School of Computer and Artificial Intelligence Zhengzhou University Zhengzhou China

Unsupervised learning utilizes unlabeled data to alleviate the reliance on large amounts of labeled data, and it has made great progress in time series anomaly detection. However, there are still some thorny issues unsolved, especially the incapability to maintain the robustness of the model when the time series dataset is contaminated. In this work, we propose an unsupervised time series anomaly detection method on data contamination, named distribution alignment autoencoder (DAAE), in which a temporal sliding reward function combines the adversarial distribution alignment to mine the contaminated data. Specifically, the dimensionality reduction and reconstruction of the time series are implemented by an autoencoder. Next, using two discriminators, the latent features obtained by dimensionality reduction are mapped to the Gaussian distribution, and the high-confidence normal time series are generated by sampling noise from the Gaussian distribution. Then, we compute the mean square error between high-confidence normal samples and the training data. Finally, we feed the error values to the temporal sliding reward function to mine the contaminated training samples. Experimental results compared with baselines on public datasets demonstrate the effectiveness of DAAE.

关键词： Training Dimensionality reduction Time series analysis Training data Mean square error methods Gaussian distribution Robustness

来源：评论

学校读者我要写书评

暂无评论

An Improved Leaky-ESN for Electricity Load Forecasting 15th

An Improved Leaky-ESN for Electricity Load Forecasting

引用

15th International Conference on Green, Pervasive, and Cloud Computing, GPC 2020

作者： Lin, Qiaoying Huang, Fangwan Yu, Zhiyong Li, Li College of Mathematics and Computer Science Fuzhou University Fuzhou China Fujian Provincial Key Laboratory of Network Computing and Intelligent Information Processing Fuzhou University Fuzhou China Key Laboratory of Spatial Data Mining and Information Sharing Ministry of Education Fuzhou China

ISBN: (纸本)9783030642426

In the recent decade, recurrent neural networks become a hot research field, with the powerful capability to capture temporal information from time series. To avoid the vanishing/exploding gradient problem caused by the gradient descent algorithm involved in most recurrent networks, echo state network (ESN) was proposed to contain a large but sparse reservoir instead of traditional hidden layers. However, the performance of ESN is very sensitive to the parameters of the reservoir. In this paper, we focus on the improvements of ESN in the background of electricity load forecasting. With the goal of effectively and efficiently computation in the context of pervasive and cloud computing, two versions of adaptive echo state network (AESN) are designed to adopt a modular control strategy to automatically adjust some parameters of the reservoir. Applying AESN on two synthetic datasets and two real world electricity datasets, experimental results demonstrate that AESN is viable. © 2020, Springer Nature Switzerland AG.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

ReactFace: Online Multiple Appropriate Facial Reaction Generation in Dyadic Interactions

arXiv

引用

arXiv 2023年

作者： Luo, Cheng Song, Siyang Xie, Weicheng Spitale, Micol Ge, Zongyuan Shen, Linlin Gunes, Hatice Computer Vision Institute College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China Guangdong Provincial Key Laboratory of Intelligent Information Processing Shenzhen University Shenzhen518060 China Department of Computer Science University of Nottingham Ningbo China Ningbo315100 China Computer Sciences University of Exeter ExeterEX4 4PY United Kingdom Department of Computer Science and Technology University of Cambridge CambridgeCB3 0FT United Kingdom Airdoc-Monash Research Centre Monash University Faculty of IT Monash University Melbourne Australia

In dyadic interaction, predicting the listener’s facial reactions is challenging as different reactions could be appropriate in response to the same speaker’s behaviour. Previous approaches predominantly treated this task as an interpolation or fitting problem, emphasizing deterministic outcomes but ignoring the diversity and uncertainty of human facial reactions. Furthermore, these methods often failed to model short-range and long-range dependencies within the interaction context, leading to issues in the synchrony and appropriateness of the generated facial reactions. To address these limitations, this paper reformulates the task as an extrapolation or prediction problem, and proposes an novel framework (called ReactFace) to generate multiple different but appropriate facial reactions from a speaker behaviour rather than merely replicating the corresponding listener facial behaviours. Our ReactFace generates multiple different but appropriate photo-realistic human facial reactions by: (i) learning an appropriate facial reaction distribution representing multiple different but appropriate facial reactions;and (ii) synchronizing the generated facial reactions with the speaker verbal and non-verbal behaviours at each time stamp, resulting in realistic 2D facial reaction sequences. Experimental results demonstrate the effectiveness of our approach in generating multiple diverse, synchronized, and appropriate facial reactions from each speaker’s behaviour. The quality of the generated facial reactions is intimately tied to the speaker’s speech and facial expressions, achieved through our novel speaker-listener interaction modules. Our code is made publicly available at https://***/lingjivoo/ReactFace. Copyright © 2023, The Authors. All rights reserved.

关键词： Extrapolation

来源：评论

学校读者我要写书评

暂无评论

A 1D Plug-and-Play Synthetic Data Deep Learning For Undersampled Magnetic Resonance Image Reconstruction

A 1D Plug-and-Play Synthetic Data Deep Learning For Undersam...

引用

IEEE International Conference on Image processing

作者： Min Xiao Zi Wang Jiefeng Guo Di Guo Xiaobo Qu Institute of Artificial Intelligence Xiamen University Xiamen China Department of Electronic Science Fujian Provincial Key Laboratory of Plasma and Magnetic Resonance Xiamen University Xiamen China Department of Microelectronics and Integrated Circuit Xiamen University Xiamen China School of Computer and Information Engineering Fujian Engineering Research Center for Medical Data Mining and Application Xiamen University of Technology Xiamen China

ISBN: (数字)9798350349399

ISBN: (纸本)9798350349405

Magnetic resonance imaging (MRI) plays a pivotal role in modern medical diagnosis yet is often hindered by the long imaging time. MRI imaging can be accelerated through undersampling, but the introduced aliasing artifacts should be removed during image reconstruction. While deep learning reconstruction methods excel at image de-aliasing, they may yield suboptimal results when training sampling settings differ from those at the time of reconstruction. To decouple from specific sampling settings, we propose using synthetic data to generate a substantial training dataset and pre-train a 1D deep denoiser. We then integrate the trained deep denoiser into the iterative reconstruction process as a replacement for the approximation operator within the deep plug-and-play framework. In vivo results indicate that the proposed method exhibits robust and visually appealing image reconstruction when there is a mismatch between the training and reconstruction undersampling settings, such as different undersampling patterns and sampling rates.

关键词： Training Deep learning In vivo Magnetic resonance imaging Magnetic resonance Training data Trajectory

来源：评论

学校读者我要写书评

暂无评论

Efficient Image Super-Resolution with Feature Interaction Weighted Hybrid Network

arXiv

引用

arXiv 2022年

作者： Li, Wenjie Li, Juncheng Gao, Guangwei Deng, Weihong Yang, Jian Qi, Guo-Jun Lin, Chia-Wen Pattern Recognition and Intelligent System Laboratory School of Artificial Intelligence Beijing University of Posts and Telecommunications Beijing100080 China Intelligent Visual Information Perception Laboratory Institute of Advanced Technology Nanjing University of Posts and Telecommunications Nanjing210046 China Key Laboratory of Artificial Intelligence Ministry of Education Shanghai200240 China Provincial Key Laboratory for Computer Information Processing Technology Soochow University Suzhou215006 China School of Communication and Information Engineering Shanghai University Shanghai200444 China School of Computer Science and Technology Nanjing University of Science and Technology Nanjing210094 China Research Center for Industries of the Future the School of Engineering Westlake University Hangzhou310024 China OPPO Research SeattleWA98101 United States Department of Electrical Engineering National Tsing Hua University Hsinchu30013 Taiwan

Lightweight image super-resolution aims to reconstruct high-resolution images from low-resolution images using low computational costs. However, existing methods result in the loss of middle-layer features due to activation functions. To minimize the impact of intermediate feature loss on reconstruction quality, we propose a Feature Interaction Weighted Hybrid Network (FIWHN), which comprises a series of Wide-residual Distillation Interaction Block (WDIB) as the backbone. Every third WDIB forms a Feature Shuffle Weighted Group (FSWG) by applying mutual information shuffle and fusion. Moreover, to mitigate the negative effects of intermediate feature loss, we introduce Wide Residual Weighting units within WDIB. These units effectively fuse features of varying levels of detail through a Wide-residual Distillation Connection (WRDC) and a Self-Calibrating Fusion (SCF). To compensate for global feature deficiencies, we incorporate a Transformer and explore a novel architecture to combine CNN and Transformer. We show that our FIWHN achieves a favorable balance between performance and efficiency through extensive experiments on low-level and high-level tasks. Codes will be available at https://***/IVIPLab/FIWHN. Copyright © 2022, The Authors. All rights reserved.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Occlusion Scene Classification via Cascade Supervised Contrastive Learning

TechRxiv

引用

TechRxiv 2022年

作者： Xu, Jianming Li, Yunfei He, Li Li, Jun The Guangdong Provincial Key Laboratory of Urbanization and Geosimulation Center of Integrated Geographic Information Analysis School of Geography and Planning Sun Yatsen University Guangzhou510275 China The School of Computer Science China University of Geosciences Wuhan430078 China The Hubei Key Laboratory of Intelligent Geo-Information Processing China University of Geosciences Wuhan430078 China The School of Automation Science and Engineering South China University of Technology Guangzhou510640 China

Recent occlusion scene classification focuses on extracting occlusion-invariant features by considering class separability, unsupervised learning, and generalization ability, separately. Considering them together is profitable to address real-world occlusion problem. In this work, we propose cascade supervised contrastive learning to accommodate those considerations simultaneously. First, the pretext task uses rectangles to occlude the original images. Second, we use the siamese residual network to extract the features of original images, and pretext task augmented images. Third, compute the cascade supervised contrastive loss of the intermediate layers and final layer of the network. By optimizing the cascade supervised contrastive loss, the network can learn pretext-task invariant features. Besides, the pretext task-invariant features can adapt to the downstream occlusion-invariant feature extraction task due to the outstanding generalization ability of contrastive learning. Finally, a multilayer perceptron is used to perform scene classification on the extracted features. The proposed method is validated on two simulated datasets and tends to achieve performance improvement compared with typical unsupervised occlusion scene classification methods. The first application of contrastive learning for remote sensing image occlusion scene classification is the innovative contribution of this paper. This method can achieve accurate remote sensing scene classification with any degree and type of occlusion. © 2022, CC BY-NC-SA.

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

High-Throughput Rice Density Estimation from Transplantation to Tillering Stages Using Deep Networks

引用

Plant Phenomics 2020年第1期2卷 255-268页

作者： Liang Liu Hao Lu Yanan Li Zhiguo Cao National Key Laboratory of Science and Technology on Multi-Spectral Information Processing School of Artificial Intelligence and AutomationHuazhong University of Science and TechnologyWuhan430074 HubeiChina The University of Adelaide AdelaideSA 5005Australia School of Computer Science and Engineering Wuhan Institute of TechnologyWuhan430205 HubeiChina

Rice density is closely related to yield estimation,growth diagnosis,cultivated area statistics,and management and damage ***,rice density estimation heavily relies on manual sampling and counting,which is inefficient and *** the prevalence of digital imagery,computer vision(CV)technology emerges as a promising alternative to automate this ***,challenges of an in-field environment,such as illumination,scale,and appearance variations,render gaps for deploying CV *** fill these gaps towards accurate rice density estimation,we propose a deep learningbased approach called the Scale-Fusion Counting Classification Network(SFC^(2)Net)that integrates several state-of-the-art computer vision *** particular,SFC^(2)Net addresses appearance and illumination changes by employing a multicolumn pretrained network and multilayer feature fusion to enhance feature *** ameliorate sample imbalance engendered by scale,SFC^(2)Net follows a recent blockwise classification *** validate SFC^(2)Net on a new rice plant counting(RPC)dataset collected from two field sites in China from 2010 to *** results show that SFC^(2)Net achieves highly accurate counting performance on the RPC dataset with a mean absolute error(MAE)of 25.51,a root mean square error(MSE)of 38.06,a relative MAE of 3.82%,and a R^(2) of 0.98,which exhibits a relative improvement of 48.2%*** over the conventional counting approach ***,SFC^(2)Net provides high-throughput processing capability,with 16.7 frames per second on 1024×1024 *** results suggest that manual rice counting can be safely replaced by SFC^(2)Net at early growth *** and models are available online at https://***/sfc2net.

关键词： illumination estimation render

来源：评论

学校读者我要写书评

暂无评论

Automatic Sleep Staging Algorithm Based on Random Forest and Hidden Markov Model

引用

computer Modeling in Engineering & Sciences 2020年第4期123卷 401-426页

作者： Junbiao Liu Duanpo Wu Zimeng Wang Xinyu Jin Fang Dong Lurong Jiang Chenyi Cai College of Information Science&Electronic Engineering Zhejiang UniversityHangzhouChina School of Communication Engineering Hangzhou Dianzi UniversityHangzhouChina Zhejiang Provincial Key Laboratory of Information Processing Communication and NetworkingHangzhouChina College of Information and Electric Engineering Zhejiang University City CollegeHangzhouChina School of Information Science and Technology Zhejiang Sci-Tech UniversityHangzhouChina Hangzhou Neuro Science and Technology Co. Ltd.HangzhouChina

In the field of medical informatics,sleep staging is a challenging and timeconsuming task undertaken by sleep *** to the new standard of the American Academy of Sleep Medicine(AASM),the stages of sleep are divided into wakefulness(W),rapid eye movement(REM)and non-rapid eye movement(NREM)which includes three sleep stages(N1,N2 and N3)that describe the depth of *** study aims to establish an automatic sleep staging algorithm based on the improved weighted random forest(WRF)and Hidden Markov Model(HMM)using only the features extracted from double-channel EEG *** WRF classification model focuses on reducing the bias of the imbalance data,while the HMM model focuses on improving the detection rate of sleep staging through the relationship between adjacent sleep *** particular,the improved weighted RF classification model can increase the recognition rate of the N1 *** addition,the method of removing features with low variance is used to select meaningful and contributing feature parameters for model *** is an innovative content of this *** sleep EEG data are first segmented into 30 s epochs,and the feature parameters of the epoch data are extracted from the double-channel by applying continuous wavelet packet transform(WPT).Each epoch is then segmented into 29 subepochs of 2 s long with 1 s overlap,and the frequency domain features and statistical features of each subepoch are *** performance of the proposed method is tested by evaluating the accuracy(AC),Kappa coefficient,Recall(R),Precision(P)and F1-score(F1).In the Sleep-EDF database,the overall AC and Kappa coefficient obtained by WRF are 93.20%and 0.890,respectively using the subject-non-independent *** the 10 sc*and 10 st*Sleep-EDF Expanded database,the overall AC and Kappa coefficient obtained by proposed method are 91.97%and 0.874,respectively using the subject-independent *** best AC and Kappa coefficient of single subject can reach 96.3

关键词： Sleep staging wavelet packet transform random forest HMM

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：