检索结果-内蒙古大学图书馆

4th International Conference on Computer Vision, Application, and Algorithm, CVAA 2024

作者： Sun, Kangjian Huo, Ju Jiang, Xiaona Gao, Yichao Cai, Chen School of Electrical Engineering and Automation Harbin Institute of Technology Harbin China National Key Laboratory of Modeling and Simulation for Complex Systems Harbin China China FAW Corporation Limited Changchun China College of Design and Engineering National University of Singapore Singapore Signal and Communication Research Institute China Academy of Railway Sciences Beijing China

ISBN: (数字)9781510687622

ISBN: (纸本)9781510687615

Low-light scenes are common in fields such as autonomous driving, which tests the robustness of intelligent systems. This paper tests the performance of object detection in low-light scenes based on the YOLOv8n model. In order to solve the constraint of limited computing power of embedded devices, the YOLOv8n model is adjusted through model pruning and knowledge distillation. The complexity of the model is reduced through these lightweight operations. While ensuring the detection performance as much as possible, the multiply-accumulate operations (MACs) and parameters of the model decreased by about 40% and 27% respectively. The proposed model is evaluated on the Exclusively Dark (ExDARK) dataset and achieves a mean average precision (mAP) value of 0.672. In addition, the proposed model is migrated to the embedded platform NVIDIA Jetson Xavier NX. The deployment process is optimized by multi-process scheduling and multi-thread scheduling. Experiments are also conducted on real data, and the results show that the proposed model can provide a feasible solution for vision-based nighttime autonomous driving. © 2025 SPIE.

关键词： Object detection Deep learning Image processing

来源：评论

学校读者我要写书评

暂无评论

NEW LIOUVILLE TYPE THEOREMS FOR 3D STEADY INCOMPRESSIBLE MHD EQUATIONS AND HALL-MHD EQUATIONS

arXiv

引用

arXiv 2025年

作者： Zhang, Zhibing School of Mathematics and Physics Key Laboratory of Modeling Simulation and Control of Complex Ecosystem in Dabie Mountains of Anhui Higher Education Institutes Anqing Normal University Anqing246133 China

In this paper, we study Liouville type results for the three-dimensional stationary incompressible MHD equations and Hall-MHD equations. By a new iteration argument, we establish Liouville type theorems if the velocity and magnetic field satisfy certain growth conditions of Lebesgue norms on the annulus. In our iteration procedure, the L2 norms of the gradients of the velocity and magnetic field and the L6 norms of the velocity and magnetic field are iterated together. The conditions imposed on the magnetic field are weaker than the velocity field in certain sense. As a consequence, we show that the velocity and magnetic field are trivial provided that they belong to some Lebesgue spaces or satisfy some decay conditions at infinity. Our results extend and improve the recent works of Chae-Lee (2024 Nonlinearity 37 095006) and Cho-Neustupa-Yang (2024 Nonlinearity 37 035007).MSC Codes 76W05, 76D03, 76D05 Copyright © 2025, The Authors. All rights reserved.

关键词： Liouville equation

来源：评论

学校读者我要写书评

暂无评论

Automatic Inverse Synthetic Aperture Radar Ship Targets Recognition using Simple Multiple Kernel Learning: Automatic ISAR Ship Targets Recognition using Simple-MKL 24

Automatic Inverse Synthetic Aperture Radar Ship Targets Reco...

引用

Proceedings of the 2024 4th International Conference on Big Data, Artificial Intelligence and Risk Management

作者： Yuan Wu Fulin Su Xue Li Peipei Zhu Xunwei Xie National Key Laboratory of Complex Aviation System Simulation Southwest China Institute of Electronic Technology Chengdu Sichuan China School of Electronics and Information Engineering Harbin Institute of Technology Harbin Heilongjiang China

ISBN: (纸本)9798400709753

In radar automatic target recognition (RATR), inverse synthetic aperture radar (ISAR) image recognition shows its advantages. Due to the limited sample size of ISAR images, support vector machine (SVM), known for its robustness in small sample classification, is often used for ISAR image recognition. For ISAR images of different targets, the single kernel SVM algorithm might lose its robustness. Therefore, this paper applies multiple kernel learning (MKL) to ISAR ship target recognition. The process begins with the preprocessing of the ISAR images to suppress Gaussian white noise. Then, principal component analysis (PCA) is employed to extract features from the ISAR images. Finally, the Simple-MKL method is used to recognize the samples. Experiments based on simulation data indicate that the method used in this paper improves the accuracy compared to other single-kernel SVM algorithms with different kernel functions.

关键词： Inverse Synthetic Aperture Radar (ISAR)

来源：评论

学校读者我要写书评

暂无评论

HRSTORY: Historical News Review Based Online Story Discovery 25

HRSTORY: Historical News Review Based Online Story Discovery

引用

Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1

作者： Renjie Zhou Haoran Ye Jian Wan Yong Liao Key Laboratory of Complex Systems Modeling and Simulation of the Ministry of Education Hangzhou Dianzi University Hangzhou Zhejiang China Zhejiang University of Science and Technology Hangzhou Zhejiang China University of Science and Technology of China Hefei Anhui China

ISBN: (纸本)9798400712456

Story discovery on news streams can help people quickly find story from vast amounts of news, improving the efficiency of information acquisition. Recent online story discovery methods encode text topics and then cluster articles into stories based on similarity. However, the results obtained by these methods are one-time, and clustered news cannot adaptively update in a continuous news stream. Additionally, the inadequate quality of article encoding and the presence of noise data deteriorate the performance of story discovery. To this end, we propose HRSTORY for online story discovery on news streams, which employs a historical news review method to enable news to continuously adapt to the latest environment in the stream data and make corrections and updates. Furthermore, HRSTORY captures better article embeddings through modeling multi-layer relational dependencies within the text. By using sentence-level noise masking, HRSTORY improves the relevance of news article representation to core topics and reduces the interference of noise data. Experiments on real news datasets show that HRSTORY outperforms the state-of-the-art algorithms in unsupervised online story discovery performance.

关键词： historical news

来源：评论

学校读者我要写书评

暂无评论

Multi-scale Dilated Attention-based Generative Adversarial Network for Infrared and Visible Image Fusion

引用

Current Optics and Photonics 2025年第2期9卷 120-129页

作者： Yang, Shunyuan Shi, Xiaoping Huo, Ju Sun, Kangjian Wang, Yu School of Astronautics Harbin Institute of Technology Harbin150001 China National Key Laboratory of Modeling and Simulation for Complex Systems Harbin150001 China School of Electrical Engineering and Automation Harbin Institute of Technology Harbin150001 China Shanghai Space Propulsion Technology Research Institute Shanghai201109 China

In the field of optics, infrared and visible images are often required for use in different situations. In particular, in the field of optical sensors, infrared and visible sensors are mainly used to obtain different band images, which are applied to improve the comprehensive information content of images and to improve the quality and availability of images. In existing infrared and visible image fusion methods, the focus is often placed on retaining the background information from visible images and the salient targets from infrared images. To address this issue, we introduce a multi-scale dilated attention module into the encoder-decoder structure of the generator. By applying dilated convolution and selfattention mechanisms, this module improves the perceptual capability of the model, thereby improving performance without increasing network complexity. This design emphasizes gradient information and detailed features in visible images. Experimental results on the public TNO dataset demonstrate that our method achieves superior visual quality and preserves the most abundant image information. Moreover, experiments on spacecraft images validate the robustness and applicability of our approach. Simultaneously, our method also provides significant technical support for the optical field. © 2025 Current Optics and Photonics.

关键词： Image coding

来源：评论

学校读者我要写书评

暂无评论

Impact of refined oceanic model resolution on the simulation of Atlantic meridional overturning circulation in LICOM3

引用

Atmospheric and Oceanic Science Letters 2025年

作者： Wang, Xiaowei Yu, Yongqiang Yu, Zipeng Lin, Pengfei National Key Laboratory of Earth System Numerical Modeling and Application Institute of Atmospheric Physics Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Earth System Numerical Simulation Facility Beijing China

The present study compares the Atlantic Meridional Overturning Circulation (AMOC) in the North Atlantic from two simulations by an oceanic general circulation model with 1° × 1° and 0.1° × 0.1° resolution, respectively, which explores the sensitivity of AMOC to the resolution. The ocean model is the latest version of the LASG/IAP Climate system Model (LICOM3), and it is forced with atmospheric data from phase 2 of the Ocean Model Intercomparison Project (OMIP2). Comprehensive comparison between the two simulations indicates that the simulated AMOC is highly sensitive to the spatial resolution. The high-resolution model (LICOM-H) simulates a deeper mixed-layer depth (MLD), due to increased surface salinity, than that in the low-resolution model (LICOM-L) in the Labrador Sea, and it also has stronger AMOC strength (the maximum climatic mean AMOC from 1958 to 2018 is 23.4 Sv for LICOM-H and 21.2 Sv for LICOM-L) and larger variability of AMOC index (2.3 Sv for LICOM-H and 1.7 Sv for LICOM-L). These differences can mainly be attributed to two main dynamic processes that benefit from high resolution. Firstly, LICOM-H can simulate much stronger boundary currents (∼0.6 m s⁻¹) than LICOM-L (0.3 m s⁻¹) in the upper ocean, which leads to saltier and warmer seawater being transported to the Labrador Sea, where it enhances deep convection. Secondly, in LICOM-H, the stronger variability of AMOC is related to the higher sensitivity of the MLD and transformation to the prescribed atmospheric forcing in the Labrador Sea. 摘要本研究关注大西洋经圈翻转环流 (AMOC) 对分辨率的敏感性, 比较了大气物理研究所自主发展的海洋环流模式LICOM (LASG/IAP Climate system Ocean Model) 不同分辨率版本对AMOC的模拟结果, 两个版本分辨率分别为1° × 1°和0.1° × 0.1°.LICOM高分辨版本模拟的AMOC强度和变率较低分辨率版本明显增加, 而且在拉布拉多海也模拟出了更深的混合层深度.在上层海洋, 高分辨率版本可以模拟比低分辨率版本强得多的边界流, 这导致更多高盐海水被输送到拉布拉多海, 增强了该区域的深对流.在高分辨率版本中, AMOC较强的变率与拉布拉多海混合层深度较高的变率和模式对大气强迫更敏感的响应有关. © 2025

关键词： Atlantic meridional overturning circulation Eddy-resolving ocean model North Atlantic Transformation

来源：评论

学校读者我要写书评

暂无评论

Imp: Highly Capable Large Multimodal Models for Mobile Devices

引用

IEEE Transactions on Multimedia 2025年

作者： Shao, Zhenwei Yu, Zhou Yu, Jun Ouyang, Xuecheng Zheng, Lihao Gai, Zhenbiao Wang, Mingyang Kuang, Zhenzhong Ding, Jiajun Hangzhou Dianzi University Key Laboratory of Complex Systems Modeling and Simulation School of Computer Science Hangzhou310005 China Harbin Institute of Technology Shenzhen Department of Computer Science and Technology Shenzhen China Hangzhou Dianzi University HDU-ITMO Joint Institute Hangzhou310005 China

By harnessing the capabilities of large language models (LLMs), recent large multimodal models (LMMs) have shown remarkable versatility in open-world multimodal understanding. Nevertheless, they are usually parameter-heavy and computation-intensive, thus hindering their applicability in resource-constrained scenarios. To this end, several lightweight LMMs have been proposed successively to maximize the capabilities under constrained scale (e.g., 3B). Despite the encouraging results achieved by these methods, most of them only focus on one or two aspects of the design space, and the key design choices that influence model capability have not yet been thoroughly investigated. In this paper, we conduct a systematic study for lightweight LMMs from the aspects of model architecture, training strategy, and training data. Based on our findings, we obtain Imp - a family of highly capable LMMs at the 2B$\sim$4B scales. Notably, our Imp-3B model steadily outperforms all the existing lightweight LMMs of similar size, and even surpasses the state-of-the-art LMMs at the 13B scale. With low-bit quantization and resolution reduction techniques, our Imp model can be deployed on a Qualcomm Snapdragon 8Gen3 mobile chip with a high inference speed of about 13 tokens/s. © 2025 IEEE.

关键词： Visual languages

来源：评论

学校读者我要写书评

暂无评论

Some new Liouville type theorems for 3D steady tropical climate model

arXiv

引用

arXiv 2025年

作者： Fang, Yan Zhang, Zhibing School of Mathematics and Physics Anqing Normal University Anqing246133 China School of Mathematics and Physics Key Laboratory of Modeling Simulation and Control of Complex Ecosystem in Dabie Mountains of Anhui Higher Education Institutes Anqing Normal University Anqing246133 China

In this paper, we study the Liouville type theorems for the stationary tropical climate model in three dimension. With the help of the delicate estimates of several integrals and an iteration argument, we establish Liouville type theorems under seventeen different assumptions. As a consequence, we show that a smooth solution is trivial provided that they belong to some Lebesgue spaces or satisfy some decay conditions at infinity. Our results extend and improve the recent work of Cho-In-Yang (2024 Appl. Math. Lett. 153 109039).MSC Codes 35B53, 35Q35, 35A02 © 2025, CC BY.

关键词： Climate models

来源：评论

学校读者我要写书评

暂无评论

NSVineCopula: R package for modeling non-stationary multivariate dependence

引用

Environmental Modelling and Software 2025年 190卷

作者： Zhang, Q. Li, Y.P. Huang, G.H. Huang, X.M. Wang, H. Wang, Z. Xu, Z.P. Wang, Y.Y. Shen, Z.Y. State Key Joint Laboratory of Environmental Simulation and Pollution Control School of Environment Beijing Normal University Beijing100875 China Department of Earth System Science Ministry of Education Key Laboratory for Earth System Modeling Institute for Global Change Studies Tsinghua University Beijing100875 China Institute for Energy Environment and Sustainable Communities University of Regina ReginaSKS4S0A2 Canada

A vine copula is a flexible method for multivariate dependence simulations that assumes stationarity. However, only a few studies have focused on non-stationarity and comprehensively developed nonstationary vine copula functions. In this study, a novel R package, NSVineCopula was developed and presented. Canonical-vine and Drawable-vine structure with 36 bivariate copula functions were considered in NSVineCopula. This package is capable of capturing non-stationary multivariate dependence, providing time-varying parameters for each bivariate copula, and quantifying the conditional probability. Notably, NSVineCopula provides a simple way for sampling non-stationary vine copulas. The capability of NSVineCopula was evaluated through two case studies: (1) agricultural drought risk assessment under compound dry-hot extreme conditions and water level prediction. The results demonstrate the advantages of NSVineCopula in non-stationary multivariate dependence analysis, and highlights the potential of NSVineCopula in many fields. Overall, NSVineCopula can provide valuable and robust functionalities for modeling nonstationary multivariate dependence. © 2025 Elsevier Ltd

关键词： Copula Multivariate dependence Non-stationarity R package Vine copula

来源：评论

学校读者我要写书评

暂无评论

A Projection Method With Modular Grad-Div Stabilization for Inductionless Magnetohydrodynamic Equations Based on Charge Conservation

引用

Mathematical Methods in the Applied Sciences 2025年

作者： Yang, Yun-Bo Xia, Yande Department of Mathematics Yunnan Normal University Kunming China Yunnan Key Laboratory of Model Analytical Mathematics and Application Yunnan Normal University Kunming China Key Laboratory of Complex System Modeling and Application for Universities in Yunnan Yunnan Normal University Kunming China

In this paper, we proposed a fully discrete projection method with modular grad-div stabilization for solving the time-dependent inductionless magnetohydrodynamic equations. The method incorporates a minimally intrusive module into the classical projection method, serving as a postprocessing step, thereby enhancing solution accuracy and improving mass conservation. Additionally, a decoupled strategy is employed to separate the magnetic and fluid field functions from the original system. Since we only need to solve several linear subsystems at each time step, the numerical solutions can be obtained efficiently. In the spatial discretization, (Formula presented.) -conforming finite element pair is used to approximate the current density and electric potential, ensuring that the discrete current density is exactly divergence-free. Therefore, the designed numerical scheme maintains the features of linearization, decoupling, unconditional energy stability, charge conservation, and improved mass conservation. The unconditional energy stability and convergence of the algorithm are rigorously analyzed and proven. Numerical results are presented to verify the robustness of the algorithm with respect to the stabilization parameters and to demonstrate the performance of the scheme, particularly in terms of its stability and accuracy. © 2025 John Wiley & Sons Ltd.

关键词： Stabilization

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

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

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：