检索结果-内蒙古大学图书馆

IEEE Global Conference on Signal and Information Processing (GlobalSIP)

作者： Jianchen Zhu Shengjie Zhao Xu Ma Gonzalo R. Arce Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai CHN School of Optics and Photonics Beijing Institute of Technology Beijing CHN Department of Electrical and Computer Engineering University of Delaware Newark Delaware USA

Compressive Sensing (CS) is a new paradigm for the efficient acquisition of signals that have sparse representation in a certain domain. Traditionally, CS has provided numerous methods for signal recovery over an orthonormal basis. However, modern applications have sparked the emergence of related methods for signals not sparse in an orthonormal basis but in some arbitrary, perhaps highly overcomplete, dictionary, particularly due to their potential to generate different kinds of sparse representation of signals. Here, we first propose the Signal space Subspace Pursuit (SSSP) algorithm, and then we derive a low bound on the number of measurements required. The algorithm has low computational complexity and provides high recovery accuracy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Localization of Deep Inpainting Using High-Pass Fully Convolutional Network

Localization of Deep Inpainting Using High-Pass Fully Convol...

引用

International Conference on Computer Vision (ICCV)

作者： Haodong Li Jiwu Huang Guangdong Key Laboratory of Intelligent Information Processing and Shenzhen Key Laboratory of Media Security National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China Shenzhen Institute of Artificial Intelligence and Robotics for Society China

ISBN: (数字)9781728148038

ISBN: (纸本)9781728148045

Image inpainting has been substantially improved with deep learning in the past years. Deep inpainting can fill image regions with plausible contents, which are not visually apparent. Although inpainting is originally designed to repair images, it can even be used for malicious manipulations, e.g., removal of specific objects. Therefore, it is necessary to identify the presence of inpainting in an image. This paper presents a method to locate the regions manipulated by deep inpainting. The proposed method employs a fully convolutional network that is based on high-pass filtered image residuals. Firstly, we analyze and observe that the inpainted regions are more distinguishable from the untouched ones in the residual domain. Hence, a high-pass pre-filtering module is designed to get image residuals for enhancing inpainting traces. Then, a feature extraction module, which learns discriminative features from image residuals, is built with four concatenated ResNet blocks. The learned feature maps are finally enlarged by an up-sampling module, so that a pixel-wise inpainting localization map is obtained. The whole network is trained end-to-end with a loss addressing the class imbalance. Extensive experimental results evaluated on both synthetic and realistic images subjected to deep inpainting have shown the effectiveness of the proposed method.

关键词： Forensics Feature extraction Convolution Kernel Machine learning Computer vision Probability

来源：评论

学校读者我要写书评

暂无评论

Parallel Building Process Management system

Parallel Building Process Management System

引用

Chinese Automation Congress (CAC)

作者： Sheng Liu Xisong Dong Gang Xiong Zhen Shen Bin Hu Xiaoyu Chen State Key Laboratory for Management and Control of Complex System Institute of Automation Chinese Academy of Sciences Beijing China Guangdong Engineering Research Center of 3D Printing and Intelligent Manufacturing Cloud Computing Center Chinese Academy of Sciences Dongguan China Beijing Engineering Research Center of Intelligent Systems and Technology Institute of Automation Chinese Academy of Sciences Beijing China Institution of Intelligent Manufacturing Qingdao Academy of Intelligent Industries Qingdao China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9781665426480

With the development of industrialized building and intelligence technology, it becomes necessary and possible to realize building process intelligence management. In recent years, digital twins and parallel management get more and more attentions in academic research and industrial applications. This paper addresses how parallel building system performs intelligence management on industrialized building process. Firstly, actual building model and artificial building model are created according to actual building system. Then business driving engine and simulation engine drive actual building model and artificial building model, respectively. Participants receive and send messages by software interfaces. Business driving engine changes the status of actual building model and sends the changes to simulation engine. Simulation engine updates the status of artificial building model and runs artificial building model with various parameters. Then it sends the parameter with the best effect to business driving engine. Business driving engine sends recommendations to participants according to the best parameter. Thus, building process intelligence management is achieved.

关键词： Automation Digital twin Buildings User interfaces Software Engines Business

来源：评论

学校读者我要写书评

暂无评论

Understanding discrepancies in soil moisture from SMAP and AMSR2: insights into performance and dry-down behavior

引用

Mapping Sciences and Remote Sensing 2025年第1期62卷

作者： Zhiqing Peng Tianjie Zhao Jiancheng Shi Lu Hu Thomas J. Jackson Michael H. Cosh Hui Lu Xiaodong Gao Jingyao Zheng Panpan Yao Qian Cui Peng Guo Peilin Song Zushuai Wei Mengjia Wang Anmin Fu a Key Laboratory of Remote Sensing and Digital Earth Aerospace Information Research Institute Chinese Academy of Sciences Beijing Chinab Department of Earth System Science Institute for Global Change Studies Tsinghua University Beijing China a Key Laboratory of Remote Sensing and Digital Earth Aerospace Information Research Institute Chinese Academy of Sciences Beijing China c National Space Science Center Chinese Academy of Sciences Beijing China d International Institute for Earth System Science Nanjing University Nanjing China e United States Department of Agriculture Agricultural Research Service Hydrology and Remote Sensing Laboratory (Retired) Beltsville MD USA f United States Department of Agriculture Agricultural Research Service Hydrology and Remote Sensing Laboratory Beltsville MD USA b Department of Earth System Science Institute for Global Change Studies Tsinghua University Beijing Chinag State Key Laboratory of Hydroscience and Engineering Tsinghua University Beijing Chinah Department of Earth System Science Tsinghua University- Xi’an Institute of Surveying and Mapping Joint Research Center for Next-Generation Smart Mapping Beijing China i Institute of Soil and Water Conservation Chinese Academy of Sciences and Ministry of Water Resources Yangling China j College of Surveying and Geo-Informatics North China University of Water Resources and Electric Power Zhengzhou China k Information Center (Hydrology Monitor and Forecast Center) Ministry of Water Resources Beijing China l College of Information Science and Engineering Shandong Agricultural University Taian China m Key Laboratory of Physical Electronics and Devices Ministry of Education Faculty of Electronic and Information Engineering Xi’an Jiaotong University Xi’an Shaanxi China n School of Artificial Intelligence Jianghan University Wuhan China o School of Geoscience and Technology Zhengzhou University Zhengzhou China p Academy of Forest Inventory and Planning National Forestry and Gr

ABSTRACTUnderstanding the dynamic changes in soil moisture (SM) is crucial for studying land–atmosphere interactions in hydrometeorology. While numerous SM datasets have been developed for passive microwave remote sensing systems operating at various frequencies, the consistency of SM dry-down patterns observed by different sensors remains uncertain. Additionally, the use of distinct algorithms across SM products complicates direct comparisons. This study addresses these issues by producing two new enhanced-resolution (approximately 10-km) SM datasets using the same multi-channel collaborative algorithm (MCCA). These datasets are retrieved from the L-band brightness temperature (Tb) from Soil Moisture Active Passive (SMAP) and the C/X/Ku-band Tb from Advanced Microwave Scanning Radiometer 2 (AMSR2), referred to as MCCA SMAP and MCCA AMSR2. The satellite-derived SM data are evaluated and compared using 40 globally distributed SM observation networks at both regional (dense network) and grid scales. The results indicate that the Pearson correlation coefficients (R) of MCCA SMAP and MCCA AMSR2 SM in 23 dense networks are 0.795 and 0.664, respectively, with unbiased root-mean-square deviation (ubRMSD) of 0.041 m3/m3and 0.048 m3/m3, respectively. Both datasets perform better at the regional scale than at the grid scale. The analysis shows that SMAP outperforms AMSR2 overall, although the sensing capabilities of both payloads decline with increasing vegetation water content (VWC). Further analysis of global SM dry-down patterns reveals minimal differences in the magnitude of SM dry-down between the two payloads, but slightly larger differences in the effective wilting point. A notable disparity was observed in SM memory, with SMAP exhibiting a significantly longer memory than AMSR2. Analysis of SM loss rates shows that SMAP has a lower loss rate compared to AMSR2, consistent with the theoretical expectation that lower-frequency observations, such as those from SMAP, obser

关键词： MCCA SMAP AMSR2 soil moisture dry-down

来源：评论

学校读者我要写书评

暂无评论

Dynamic target tracking based on corner enhancement with Markov decision process

Dynamic target tracking based on corner enhancement with Mar...

引用

作者： Zuo, Guoyu Du, Tingting Ma, Lei Faculty of Information Technology Beijing University of Technology Beijing China Beijing Key Laboratory of Computing Intelligence and Intelligent System Beijing China

The tracking–learning–detection (TLD) algorithm applied in the home environment can effectively improve the tracking robustness. However, it has the problems of single target tracking and poor selection of feature points. This study proposed a dynamic target tracking method based on corner enhancement with Markov decision process (MDP) model. The MDP target tracking method is adopted to change a multi-target tracking problem into a strategy problem based on MDP model, in which one MDP model represents the life cycle of a target, and multiple targets are represented by multiple MDP models. In the tracking process, the strong corners generated by the Shi-Tomasi corner method are used to replace the feature points generated by the traditional TLD algorithm at intermediate intervals, which makes the target feature points more stable during the tracking process. The similarity function learning for data association is equivalent to the learning of the MDP strategy, in which the reinforcement learning method is used and has double advantages of both online and offline learning. The tracking experiments with different data sets are performed, and the results show that dynamic target tracking algorithm based on the corner enhancement with MDP has both good tracking performance and good anti-interference capability. © 2018 Institution of Engineering and technology. All rights reserved.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Joint object contour points and semantics for instance segmentation

arXiv

引用

arXiv 2020年

作者： Zhang, Wenchao Fu, Chong Zhu, Mai Cao, Lin Tie, Ming Sham, Chiu-Wing School of Computer Science and Engineering Northeastern University Shenyang110819 China Engineering Research Center of Security Technology of Complex Network System Ministry of Education China Key Laboratory of Intelligent Computing in Medical Image Ministry of Education Northeastern University Shenyang110819 China School of Information and Communication Engineering Beijing Information Science and Technology University Beijing100101 China Science and Technology on Space Physics Laboratory Beijing100076 China School of Computer Science University of Auckland New Zealand

The attributes of object contours has great significance for instance segmentation task. However, most of the current popular deep neural networks do not pay much attention to the object edge information. Inspired by the human annotation process when making instance segmentation datasets, in this paper, we propose Mask Point R-CNN aiming at promoting the neural network’s attention to the object boundary. Specifically, we innovatively extend the original human keypoint detection task to the contour point detection of any object. Based on this analogy, we present an contour point detection auxiliary task to Mask R-CNN, which can boost the gradient flow between different tasks by effectively using feature fusion strategies and multi-task joint training. As a consequence, the model will be more sensitive to the edges of the object and can capture more geometric features. Quantitatively, the experimental results show that our approach outperforms vanilla Mask R-CNN by 3.8% on Cityscapes dataset and 0.8% on COCO dataset. Copyright © 2020, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-modal remote sensory learning for multi-objects over autonomous devices

引用

Frontiers in bioengineering and biotechnology 2025年 13卷 1430222页

作者： Aysha Naseer Naif Almudawi Hanan Aljuaid Abdulwahab Alazeb Yahay AlQahtani Asaad Algarni Ahmad Jalal Hui Liu Department of Computer Science Air University Islamabad Pakistan. Department of Information Systems College of Computer and Information Sciences Princess Nourah Bint Abdulrahman University Riyadh Saudi Arabia. Department of Computer Science College of Computer Science and Information System Najran University Najran Saudi Arabia. Department of Informatics and Computer Systems King Khalid University Abha Saudi Arabia. Department of Computer Sciences Faculty of Computing and Information Technology Northern Border University Rafha Saudi Arabia. Department of Computer Science and Engineering College of Informatics Korea University Seoul South Korea. Guodian Nanjing Automation Co. Ltd. Nanjing China. Jiangsu Key Laboratory of Intelligent Medical Image Computing School of Future Technology Nanjing University of Information Science and technology Nanjing China. Cognitive Systems Lab University of Bremen Bremen Germany.

Introduction:There has been an increasing focus on object segmentation within remote sensing images in recent years due to advancements in remote sensing technology and the growing significance of these images in both military and civilian realms. In these situations, it is critical to accurately and quickly identify a wide variety of objects. In many computer vision applications, scene recognition in aerial-based remote sensing imagery presents a common issue. Method:However, several challenging elements make this work especially difficult: (i) Different objects have different pixel densities; (ii) objects are not evenly distributed in remote sensing images; (iii) objects can appear differently depending on viewing angle and lighting conditions; and (iv) there are fluctuations in the number of objects, even the same type, in remote sensing images. Using a synergistic combination of Markov Random Field (MRF) for accurate labeling and Alex Net model for robust scene recognition, this work presents a novel method for the identification of remote sensing objects. During the labeling step, the use of MRF guarantees precise spatial contextual modeling, which improves comprehension of intricate interactions between nearby aerial objects. By simultaneously using deep learning model, the incorporation of Alex Net in the following classification phase enhances the model's capacity to identify complex patterns in aerial images and adapt to a variety of object attributes. Results:Experiments show that our method performs better than others in terms of classification accuracy and generalization, indicating its efficacy analysis on benchmark datasets such as UC Merced Land Use and AID. Discussion:Several performance measures were calculated to assess the efficacy of the suggested technique, including accuracy, precision, recall, error, and F1-Score. The assessment findings show a remarkable recognition rate of around 97.90% and 98.90%, on the AID and the UC Merced Land datasets,

关键词： autonomous devices deep learning computer vision likelihood estimation multi-modal multi-objects posterior probability remote sensing scene analysis

来源：评论

学校读者我要写书评

暂无评论

BuildSenSys: Reusing building sensing data for traffic prediction with cross-domain learning

arXiv

引用

arXiv 2020年

作者： Fan, Xiaochen Xiang, Chaocan Chen, Chao Yang, Panlong Gong, Liangyi Song, Xudong Nanda, Priyadarsi He, Xiangjian School of Electrical and Data Engineering Faculty of Engineering and Information Technology University of Technology SydneyNSW2007 Australia Key Laboratory of Dependable Service Computing in Cyber Physical Society Chongqing University Ministry of Education China College of Computer Science Chongqing University Chongqing400044 China School of Computer Science and Technology University of Science and Technology of China Hefei Anhui230026 China School of Software BNRist Tsinghua University Beijing100084 China

With the rapid development of smart cities, smart buildings are generating a massive amount of building sensing data by the equipped sensors. Indeed, building sensing data provides a promising way to enrich a series of data-demanding and cost-expensive urban mobile applications. In this paper, as a preliminary exploration, we study how to reuse building sensing data to predict traffic volume on nearby roads. Compared with existing studies, reusing building sensing data has considerable merits of cost-efficiency and high-reliability. Nevertheless, it is non-trivial to achieve accurate prediction on such cross-domain data with two major challenges. First, relationships between building sensing data and traffic data are not unknown as prior, and the spatio-temporal complexities impose more difficulties to uncover the underlying reasons behind the above relationships. Second, it is even more daunting to accurately predict traffic volume with dynamic building-traffic correlations, which are cross-domain, non-linear, and time-varying. To address the above challenges, we design and implement BuildSenSys, a first-of-its-kind system for nearby traffic volume prediction by reusing building sensing data. Our work consists of two parts, i.e., Correlation Analysis and Cross-domain Learning. First, we conduct a comprehensive building-traffic analysis based on multi-source datasets, disclosing how and why building sensing data is correlated with nearby traffic volume. Second, we propose a novel recurrent neural network for traffic volume prediction based on cross-domain learning with two attention mechanisms. Specifically, a cross-domain attention mechanism captures the building-traffic correlations and adaptively extracts the most relevant building sensing data at each predicting step. Then, a temporal attention mechanism is employed to model the temporal dependencies of data across historical time intervals. The extensive experimental studies demonstrate that BuildSenSys outperf

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Fog-Haze Transformation and Driving Factors in Coastal Region of the Yangtze River Delta

SSRN

引用

SSRN 2022年

作者： Lyu, Rui Gao, Wei Peng, Yarong Jia, Hailing Qian, Yijie He, Qianshan Cheng, TianTao Yu, Xingna Zhao, Gang Department of Environmental Science and Engineering Fudan University Shanghai200438 China Shanghai Meteorological Bureau Shanghai20030 China Institute for Meteorology Universität Leipzig Leipzig Germany Department of Atmospheric and Oceanic Sciences Institute of Atmospheric Sciences Fudan University Shanghai200438 China National Observations and Research Station for Wetland Ecosystems of the Yangtze Estuary Shanghai200438 China Shanghai Qi Zhi Institute Shanghai200232 China Innovation Center of Ocean and Atmosphere System Zhuhai Fudan Innovation Research Institute Zhuhai518057 China Key Laboratory for Aerosol-Cloud-Precipition of China Meteorological Administration Nanjing University of Information Science and Technology Nanjing210044 China Yunnan Meteorological Service Center Kunming650034 China

Low-visibility events (LVEs) are the damaging weathers that cause harm even disasters to human, closely related to anthropogenic pollution, and have negative impacts on traffic, air quality, human health, environment and etc. The 2-months wintertime continuous measurements conducted in Chongming Island of Shanghai were employed to examine LVEs on hour time scale and characterize their evolution and key drivers, particularly of fog-haze transformation. LVEs account for 38% of hours during the entire campaign, of which mist 14%, haze 6% and fog 5%, surprisingly fog-haze coexistence as high as about 13%. LVEs generally interdepend and transform each other, and their formation and evolution are the results of the synergy impacts of meteorological conditions and particulate matter. Fog and mist mostly occur from midnight to early morning, whereas haze mostly occurs in the daytime. Fog turning into haze often occurs after sunrise, while haze turning into fog often occurs after sunset. Temperature difference is the most critical meteorological driving factor of LVEs evolution. The radiative cooling on surface is more likely to improve ambient saturation than humidifying in the short term to promote fog and mist formation. During fog and mist events, under superior humidity conditions, particles exhibit relatively lower hygroscopicity and intense internal mixing due to fog droplet removal and aerosol heterogeneous and aqueous reactions. High CCN concentration accelerates the increase of fog droplet number and liquid water content. Overall, the conditions of high humidity and particle loadings are beneficial to conduce LVEs, while the conditions of sufficient water vapor at low particle level and sufficient particles at low humidity level can also trigger LVEs. This study provides insights into LVEs classification, evolution scheme and aerosol roles from a micro perspective, and the results will be helpful to improve the accuracy of refined forecast to local radiative fog an

关键词： Aerosols

来源：评论

学校读者我要写书评

暂无评论

Toward the First SDN Programming Capacity Theorem on Realizing High-Level Programs on Low-Level Datapaths

Toward the First SDN Programming Capacity Theorem on Realizi...

引用

2018 IEEE Conference on Computer Communications, INFOCOM 2018

作者： Leet, Christopher Wang, Xin Richard Yang, Y. Aspnes, James Department of Computer Science Yale University United States Department of Computer Science Tongji University China Key Laboratory of Embedded System and Service Computing Ministry of Education China

ISBN: (纸本)9781538641286

High-Ievel programming and programmable data paths are two key capabilities of software-defined networking (SDN). A fundamental problem linking these two capabilities is whether a given high-level SDN program can be realized onto a given low-level SDN datapath structure. Considering all high-level programs that can be realized onto a given datapath as the programming capacity of the datapath, we refer to this problem as the SDN data path programming capacity problem. In this paper, we conduct the first study on the SDN datapath programming capacity problem, in the general setting of high-level, datapath oblivious, algorithmic SDN programs and state-of-art multi-table SDN data path pipelines. In particular, considering datapath-oblivious SDN programs as computations and datapath pipelines as computation capabilities, we introduce a novel framework called SDN characterization junctions, to map both SDN programs and datapaths into a unifying space, deriving the first rigorous result on SDN datapath programming capacity. We not only prove our results but also conduct realistic evaluations to demonstrate the tightness of our analysis. © 2018 IEEE.

关键词： Computation theory

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：