检索结果-内蒙古大学图书馆

2021 3rd International Forum on Geoscience and Oceanography, IFGO 2021

作者： Qiao, Yulin Mohammad, Mostofa Zaman Li, Yi Liu, Xiaobo Cai, Zhihua Wuhan430074 China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan430074 China Wuhan430078 China Hubei Key Laboratory of Intelligent Geo-Information Processing China University of Geosciences Wuhan430078 China

Sample generation is an effective method to improve the performance of hyperspectral image classification by generating virtual samples for training sample expansion in the training process of classification. However, there are some defects existing in the previous sample generation methods including the lack of spatial information, the redundant generation and the damage of the original spectral components. In this paper, we propose conditional band selection generative adversarial net, named CBS-GAN, to handle this problem. Firstly, the band selection net of CBS-GAN is utilized to avoid redundant bands and keep original spectral information, then the generation net of CBS-GAN generates the spatial-spectral data blocks by selected bands for sample generation. The experiments of classification are also used to demonstrate the availability of virtual samples generated by our method. © Published under licence by IOP Publishing Ltd.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Simplified outlier detection for improving the robustness of a fuzzy model

引用

Science China(Information Sciences) 2020年第4期63卷 218-220页

作者： Yali JIN Weihua CAO Min WU Yan YUAN School of Automation China University of Geosciences Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems China University of Geosciences

Dear editor,Fuzzy models have been widely employed in various fields, such as complex industrial process modeling, classification, function approximation, and time series prediction. Extracting fuzzy rules is very important as they determine the capabilities of the fuzzy model. In recent years, data-based methods such as fuzzy clustering, genetic algorithm, and neural networks [1–3], have been pro-

关键词： Simplified outlier detection for improving the robustness of a fuzzy model Gaussian

来源：评论

学校读者我要写书评

暂无评论

Multisensor fusion estimation of nonlinear systems with intermittent observations and heavy-tailed noises

引用

Science China(Information Sciences) 2022年第09期65卷 173-190页

作者： Bo XIAO Q.M.Jonathan WU Liping YAN School of Artificial Intelligence Beijing University of Posts and Telecommunications Department of Electrical and Computer Engineering University of Windsor Key Laboratory of Intelligent Control and Decision of Complex Systems School of AutomationBeijing Institute of Technology

Inspired by the robust student t-distribution based nonlinear filter(RSTNF), a student tdistribution and unscented transform(UT) based filter for state estimation of heavy-tailed nonlinear dynamic systems, a modified RSTNF for intermittent observations is derived. The fusion estimation for nonlinear multisensor systems with intermittent observations and heavy-tailed measurement and process noises is *** this work, the centralized fusion, the sequential fusion, and the na¨?ve distributed fusion algorithms are presented, respectively. Theoretical analysis shows that the presented algorithms are effective, which are the efficient extension of the classical unscented Kalman filter(UKF) or the cubature Kalman filter(CKF) based algorithms with Gaussian noises. Simulation results show that the presented algorithms are effective and feasible.

关键词： state fusion estimation nonlinear systems heavy-tailed noise intermittent observations multivariate t-distribution

来源：评论

学校读者我要写书评

暂无评论

Adapting Human Mesh Recovery with Vision-Language Feedback

arXiv

引用

arXiv 2025年

作者： Xu, Chongyang Huang, Buzhen Zhang, Chengfang Feng, Ziliang Wang, Yangang College of Computer Science Sichuan University Chengdu610065 China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education The School of Automation Southeast University Nanjing210096 China Intelligent Policing Key Laboratory of Sichuan Province Sichuan Police College Luzhou646000 China

Human mesh recovery can be approached using either regression-based or optimization-based methods. Regression models achieve high pose accuracy but struggle with model-to-image alignment due to the lack of explicit 2D-3D correspondences. In contrast, optimization-based methods align 3D models to 2D observations but are prone to local minima and depth ambiguity. In this work, we leverage large vision-language models (VLMs) to generate interactive body part descriptions, which serve as implicit constraints to enhance 3D perception and limit the optimization space. Specifically, we formulate monocular human mesh recovery as a distribution adaptation task by integrating both 2D observations and language descriptions. To bridge the gap between text and 3D pose signals, we first train a text encoder and a pose VQ-VAE, aligning texts to body poses in a shared latent space using contrastive learning. Subsequently, we employ a diffusion-based framework to refine the initial parameters guided by gradients derived from both 2D observations and text descriptions. Finally, the model can produce poses with accurate 3D perception and image consistency. Experimental results on multiple benchmarks validate its effectiveness. The code will be made publicly available. © 2025, CC BY.

关键词： Mesh generation

来源：评论

学校读者我要写书评

暂无评论

Wearable ubiquitous energy system

引用

Science China(Information Sciences) 2021年第2期64卷 237-239页

作者： Fang DENG Ning DING Ziman YE Yeyun CAI Jie CHEN School of Automation Beijing Institute of Technology Key Laboratory of Intelligent Control and Decision of Complex Systems Beijing Institute of Technology

In recent years, owing to the growing maturity of Internet of Things technology and mobile communication technology, intelligent wearable devices have developed *** have been widely applied in aviation, medicine, military, entertainment and other fields. Through intelligent wearable devices, people can communicate efficiently, dynamically perceive external environments, and monitor the body's vital signs.

关键词： Information systems and Communication Service

来源：评论

学校读者我要写书评

暂无评论

Mdbfusion: A Visible And Infrared Image Fusion Framework Capable For Motion Deblurring

Mdbfusion: A Visible And Infrared Image Fusion Framework Cap...

引用

IEEE International Conference on Image Processing

作者： Jun Chen Wei Yu Xin Tian Jun Huang Jiayi Ma School of Automation China University of Geosciences Wuhan China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Electronic Information School Wuhan University Wuhan China

ISBN: (数字)9798350349399

ISBN: (纸本)9798350349405

As a vital technology in the field of image processing, numerous visible and infrared fusion methods have emerged with the aim of generating a single image containing salient targets and abundant details. However, owing to the influence of target motion or camera shake, source images may suffer from motion blur, resulting in blurred edges and unrecognizable objects. To address this issue, this paper bridges the gap between deblurring and fusion tasks and proposes a joint motion deblurring and fusion network for visible and infrared image (MDbFusion). On the one hand, we innovatively merge motion deblurring task into network design, which effectively ensures the capacity of MDbFusion to process images in the extreme condition. On the other hand, an Adaptive Weight Module (AWM) is designed to calculate contribution between visible and infrared features, which solves the channel contrast between two tasks, greatly reducing the complexity of network. Extensive experiments demonstrate that the MDb-Fusion outperforms state-of-the-art (SOTA) fusion algorithms in terms of preserving texture details and quantitative metrics. Furthermore, we also compare with a two-step fusion strategy that first deblurring then fusion, both with SOTA methods. The results reveal superiority of our framework in coupling and reciprocity between two tasks.

关键词： Measurement Couplings Adaptive systems Image edge detection Cameras complexity theory Decoding

来源：评论

学校读者我要写书评

暂无评论

Parallel-Data-Based Social Evolution Modeling

引用

Tsinghua Science and Technology 2021年第6期26卷 878-885页

作者： Weishan Zhang Zhaoxiang Hou Xiao Wang Zhidong Xu Xin Liu Fei-Yue Wang China University of Petroleum Qingdao 266580China Qingdao Academy of Intelligent Industry Qingdao 266111China State Key Laboratory of Management and Control for Complex Systems Institute of AutomationChinese Academy of SciencesBeijing 100049China Institute of National Security National Defense UniversityBeijing 100081China

Abnormal or drastic changes in the natural environment may lead to unexpected events,such as tsunamis and earthquakes,which are becoming a major threat to national ***,no effective assessment approach can deduce a situation and determine the optimal response strategy when a natural disaster *** this study,we propose a social evolution modeling approach and construct a deduction model for self-playing,self-learning,and self-upgrading on the basis of the idea of parallel data and reinforcement *** proposed approach can evaluate the impact of an event,deduce the situation,and provide optimal strategies for *** the breakage of a submarine cable caused by earthquake as an example,we find that the proposed modeling approach can obtain a higher reward compared with other existing methods.

关键词： paral el data reinforcement learning decision-making

来源：评论

学校读者我要写书评

暂无评论

Multiscale Feature Fusion Module for Prohibited Item Detection

Multiscale Feature Fusion Module for Prohibited Item Detecti...

引用

6th International Conference on intelligent computing and Signal Processing (ICSP)

作者： Ruiqi Xu Chirui Zhang School of Automation China University of Geosciences Wuhan Hubei China Hubei key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan Hubei China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan Hubei China

ISBN: (数字)9798350376548

ISBN: (纸本)9798350376555

X-ray security check is a prevalent placement measure that plays a vital role in ensuring public safety. However, detecting prohibited items poses a complex challenge due to their variable sizes within X-ray images. In this study, we tackle this issue by curating a novel dataset through the integration of open-source data and introducing a Multiscale Feature Fusion Module. This module facilitates robust detection by amalgamating features across different scales. Extensive experiments verified that the superiority of our method compared to existing methods on the prohibited items test set constructed in this paper demonstrates extremely well the practical effectiveness of our proposed method.

关键词： Signal processing Feature extraction Public security Security Data mining X-ray imaging

来源：评论

学校读者我要写书评

暂无评论

MixSA: Training-free Reference-based Sketch Extraction via Mixture-of-Self-Attention

引用

IEEE Transactions on Visualization and Computer Graphics 2024年 PP卷 PP页

作者： Yang, Rui Wu, Xiaojun He, Shengfeng Shaanxi Normal University School of Computer Sciences Xi'an China Shaanxi Normal University School of Computer Science The Key Laboratory of Intelligent Computing and Service Technology for Folk Song The Ministry of Culture and Tourism Xi'an710119 China Singapore Management University School of Computing and Information Systems Singapore Singapore

Current sketch extraction methods either require extensive training or fail to capture a wide range of artistic styles, limiting their practical applicability and versatility. We introduce Mixture-of-Self-Attention (MixSA), a training-free sketch extraction method that leverages strong diffusion priors for enhanced sketch perception. At its core, MixSA employs a mixture-of-self-attention technique, which manipulates self-attention layers by substituting the keys and values with those from reference sketches. This allows for the seamless integration of brushstroke elements into initial outline images, offering precise control over texture density and enabling interpolation between styles to create novel, unseen styles. By aligning brushstroke styles with the texture and contours of colored images, particularly in late decoder layers handling local textures, MixSA addresses the common issue of color averaging by adjusting initial outlines. Evaluated with various perceptual metrics, MixSA demonstrates superior performance in sketch quality, flexibility, and applicability. This approach not only overcomes the limitations of existing methods but also empowers users to generate diverse, high-fidelity sketches that more accurately reflect a wide range of artistic expressions. © 1995-2012 IEEE.

关键词： Image texture

来源：评论

学校读者我要写书评

暂无评论

Improved Shufflenet V2 Network with Attention for Speech Emotion Recognition

SSRN

引用

SSRN 2024年

作者： Udeh, Chinonso Paschal Chen, Luefeng Du, Sheng Liu, Yulong Li, Min Wu, Min School of Automation China University of Geosciences Wuhan430074 China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan430074 China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan430074 China

This paper introduces a novel algorithm for recognizing speech emotions, utilizing an enhanced version of the ShuffleNet V2 network. Previous studies typically analyze speech signals solely from the time or frequency domain, resulting in the loss of detailed information. Our algorithm initially converts speech data into an acoustic spectrogram to address this limitation, preserving rich original information across both time and frequency domains. This approach optimizes feature extraction by leveraging the capabilities of deep convolutional networks to capture local changes in images. Furthermore, we integrated the attention mechanism from ECA-Net into the ShuffleNet V2 model. This incorporation enhances the extraction of significant features, amplifying the expression of features strongly correlated with the emotional state conveyed in speech while suppressing irrelevant ones. In addition to these enhancements, we replace the ReLU activation function with Hardswish to deepen the neural network, thereby improving feature information without compromising model accuracy. Finally, the training stage incorporates the hot restart Cosine Annealing Learning Rate to optimize the network further. Simulation experiments are conducted on the enterface'05 dataset to evaluate the accuracy of the proposed algorithm. The experimental results show that the recognition rate of the method is higher than the existing algorithms. © 2024, The Authors. All rights reserved.

关键词： Human robot interaction

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：