检索结果-内蒙古大学图书馆

IEEE Visual Communications and Image Processing (VCIP)

作者： Xiuhuan Zang Yunhui Shi Jin Wang Wenpeng Ding Baocai Yin Beijing Advanced Innovation Center for Future Internet Technology Beijing Key Laboratory of Multimedia and Intelligent Software Technology School of Software Engineering Beijing University of Technology Beijing China College of Computer Science and Technology Dalian University of Technology Dalian China

ISBN: (纸本)9781509053179

Compressive light field photography has gained a significant research interest during the past years. One important problem in light field images is the effective representation of the structures. In this paper, based on collaborative projection and sparse dictionary, we propose a light field reconstruction optimization scheme for compressive light field photography. By utilizing the special structural feature of high correlations among light field views, an optimized structured dictionary is proposed, which is more memory and time efficient compared with general sparse dictionary. Furthermore, we propose the corresponding structured dictionary learning model and efficient algorithm to train the optimized dictionary with fixed projection. Simulation results on light field data set demonstrate the effectiveness of our optimizing scheme in terms of reconstruction quality compared with general sparse dictionary.

关键词： Dictionaries Image reconstruction Training Collaboration Photography Algorithm design and analysis Compressed sensing

来源：评论

学校读者我要写书评

暂无评论

Low-Rank Representation based Traffic Data Completion Method

Low-Rank Representation based Traffic Data Completion Method

引用

International Joint Conference on Neural Networks

作者： Rong Du Yong Zhang Boyue Wang Hao Liu Guanglei Qi Baocai Yin Beijing Key Laboratory of Multimedia and Intelligent Software Technology College of Metropolitan Transportation Beijing University of Technology 100124 China Beijing Transportation Information Center China

ISBN: (纸本)9781509006212

intelligent Transportation Systems (ITS) plays a significant role in the traffic management, i.e. traffic jam prediction, route guidance. Due to the hardware failure or data transformation failure, some traffic observation data may be occasionally missed, which seriously affect intelligent transportation information service. So, the completion of traffic observation data has now become an issue that requires to be concerned and solved. By analyzing the traffic history data, we find that traffic data tend to have strong spatio-temporal correlation. Considering this feature, we propose a new low-rank representation based traffic data completion method. To further enhance the local correlation, we introduce an ordered regulation into our proposed method. We also give an efficient solution to our proposed methods. In order to verify the performance of our methods, some traffic data completion experiments are conducted on the beijing metropolitan road speed dataset and the capital airport highway microwave dataset. Experimental results show that the proposed methods are superior to other state-of-the-art traffic data completion methods.

关键词： Correlation Roads Principal component analysis Sparse matrices Data models Bayes methods Traffic data Sparse matrix data models Bayes methods Road Completion principal components analysis Traffic management intelligent transportation systems Hardware Failure

来源：评论

学校读者我要写书评

暂无评论

Cross-media Hot Topic Auto-tracking Model Based on Semantics and Temporal Context

引用

Chinese Journal of Electronics 2015年第3期24卷 529-534页

作者： LIANG Meiyu DU Junping ZHOU Yipeng Beijing Key Lab of Intelligent Telecommunication Software and Multimedia School of Computer ScienceBeijing University of Posts and Telecommunications School of Computer and Information Engineering Beijing Technology and Business University

As the Internet multimedia information grows explosively, seeking an automatic technology to realize the effective organization and management of crossmedia emergency information is significantly necessary. A novel cross-media hot topic auto-tracking model based on semantics and temporal context is proposed in this paper. According to the semantic correlations of cross-media information, we learn the image visual semantics by the text semantics based on the Latent Dirichlet Allocation probability model, and establish the unified cross-media information description on the same semantic level. Also a semantics-based two-step feature dimension-reduction scheme is proposed to establish the efficient semantic feature space. The self-adaptive learning of topic model is realized to track the dynamic changes in the topic. Experimental results demonstrate that the proposed method outperforms the existing methods, which further improves the effect of hot topic auto-tracking.

关键词： Cross-media data mining Topic autotracking Semantic model Latent dirichlet allocation(LDA) Temporal context

来源：评论

学校读者我要写书评

暂无评论

ACCELERATE CONVOLUTIONAL NEURAL NETWORKS FOR BINARY CLASSIFICATION VIA CASCADING COST-SENSITIVE FEATURE

ACCELERATE CONVOLUTIONAL NEURAL NETWORKS FOR BINARY CLASSIFI...

引用

IEEE International Conference on Image Processing

作者： Junbiao Pang Huihuang Lin Li Su Chunjie Zhang Weigang Zhang Lijuan Duan Qingming Huang Baocai Yin Beijing Key Laboratory of Multimedia and Intelligent Software Technology China School of Computer and Control Engineering University of Chinese Academy of Sciences China School of Computer Science and Technology Harbin Institute of Technology at Weihai China College of Computer Science and Technology Beijing University of Technology China

ISBN: (纸本)9781467399623

Convolutional Neural Networks (CNNs) have delivered impressive state-of-the-art performances for many vision tasks, while the computation costs of these networks during test-time are notorious. Empirical results have discovered that CNNs have learned the redundant representations both within and across different layers. When CNNs are applied for binary classification, we investigate a method to exploit this redundancy across layers, and construct a cascade of classifiers which explicitly balances classification accuracy and hierarchical feature extraction costs. Our method cost-sensitively selects feature points across several layers from trained networks and embeds non-expensive yet discriminative features into a cascade. Experiments on binary classification demonstrate that our framework leads to drastic test-time improvements, e.g., possible 47.2× speedup for TRECVID upper body detection, 2.82× speedup for Pascal VOC2007 People detection, 3.72× for INRIA Person detection with less than 0.5% drop in accuracies of the original networks.

关键词： Neural networks Convolution Acceleration Optimization Upper bound computers computer science

来源：评论

学校读者我要写书评

暂无评论

Emergency online attention & psychological distance under risk 2

Emergency online attention & psychological distance under ri...

引用

2nd ACM SIGSPATIAL International Workshop on the Use of GIS in Emergency Management, EM-GIS 2016

作者： Deng, Qing Li, Lihua Liu, Yi Deng, Xiaolong Zhang, Hui Department of Engineering Physics Tsinghua University China Counter-Terrorism College People's Public Security University of China China Public Order School People's Public Security University of China China Beijing Key Laboratory of Intelligent Telecommunications Software and Multimedia Beijing University of Posts and Telecommunications China

ISBN: (纸本)9781450345804

Risk communication is an effective means of emergency management. Online information plays an important role in risk communication, especially in the Big Data Era. Citizens' psychology to send and receive information determines their online behavior when they face risk. Using the Tianjin Port Explosions as an example, multiple linear regression analysis is used to untangle the relationship between online attention and psychology behavior under risk. Citizens' online attention is estimated by social media data collection from the Sina Weibo. Psychology behavior is quantified by psychological distance which consists of four dimensions: spatial distance, temporal distance, social distance and probability. The regression model is built via SPSS 20.0 and the obtained result is matched with actual situation. It indicates that online attention is negatively correlated to spatial distance, temporal distance and social distance while positively correlated to probability of the event. It also shows that citizens' online attention under risk is positively correlated to their online attention under normal circumstance. Based on the regression model, citizens' attention and response to emergency are easy to be assessed. CCS Concepts Information systems → Data analytics Mathematics of computing → Regression analysis Mathematics of computing → Exploratory data analysis. © 2016 ACM.

关键词： Risk management

来源：评论

学校读者我要写书评

暂无评论

A Real-Time People Number Detection Algorithm of Scenic Spot Based on Density Center Clustering

A Real-Time People Number Detection Algorithm of Scenic Spot...

引用

World Congress on intelligent Control and Automation

作者： Wen Yaomin Du Junping Liang Meiyu Fan Dan JangMyung Lee Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia School of Computer Science Beijing University of Posts and Telecommunications Department of Electronics Engineering Pusan National University

ISBN: (纸本)9781467384155

In recent years, accidents occur frequently because the people number of scenic spots is lack of control. In this paper, a novel real-time people number detection algorithm of scenic spot based on density center clustering (DPBC) is proposed. Taking account of the complexity of the scenic environment, we use the Gauss mixture model (GMM) to suppress the background interference, extract the feature points from the crowd and cluster these feature points by applying a clustering algorithm based on density center. Then we establish a training set and estimate the people number by using the support vector regression model (SVR). We implement the proposed algorithm based on Opencv and use PETS dataset to validate the effectiveness of our proposed algorithm. Experimental results demonstrate that compared with Aibiol algorithm and Conte algorithm, our proposed algorithm improved the accuracy for estimating the people number in a scenic spot.

关键词： Feature extraction Clustering algorithms Algorithm design and analysis Interference Communications technology Estimation Prediction algorithms

来源：评论

学校读者我要写书评

暂无评论

ABNORMAL EVENT DETECTION BASED ON SOCIAL FORCE MODEL COMBINED WITH CROWD VIOLENT FLOW

ABNORMAL EVENT DETECTION BASED ON SOCIAL FORCE MODEL COMBINE...

引用

2016 4th IEEE International Conference on Cloud Computing and Intelligence Systems（IEEE CCIS2016）

作者： Yaomin Wen Junping Du Jang Myung Lee Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia School of Computer ScienceBeijing University of Posts and Telecommunications Department of Electronics Engineering Pusan National University

ISBN: (纸本)9781509012572

In this study, we use the social force model(SFM) combined with the violent flows descriptor(vif) to detect abnormal events. Our method extracts the local density to calculate the parameters of SFM and the vif to represent the crowd stability status. The Latent Dirichlet Allocation(LDA) model is established based on the bag-of-words model combined with the temporal and spatial features of visual words, and then the maximum likelihood function is used to identify abnormal events.

关键词： social force model Violent flow LDA model maximum likelihood

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of Tourism Activity Recognition and Discovery System

Design and Implementation of Tourism Activity Recognition an...

引用

World Congress on intelligent Control and Automation

作者： Yuan Yifan Du Junping Fan Dan JangMyung Lee Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia School of Computer Science Beijing University of Posts and Telecommunications Department of Electronics Engineering Pusan National University

ISBN: (纸本)9781467384155

With the rapid development of technologies, such as cloud computing, big data and so forth, tourism big data has drawn the attention of many scholars. How to make full use of a large amount of raw data from tourism activity, and how to recognize and discover tourism activity quickly, accurately and conveniently from information which is accumulated over a long period have become a new direction of tourism data application. In this paper, text analysis, text classification and other relevant technologies are combined based on text mining. Furthermore, a tourism activity recognition and discovery system is designed and implemented, which has achieved the recognition and discovery of tourism activity from a large amount of information. The research results could contribute to recommending and predicting tourism activity in China.

关键词： tourism technology Development big data text mining Outcomes Research

来源：评论

学校读者我要写书评

暂无评论

TOURISM ACTIVITY RECOGNITION AND DISCOVERY BASED ON IMPROVED LDA MODEL

TOURISM ACTIVITY RECOGNITION AND DISCOVERY BASED ON IMPROVED...

引用

2016 4th IEEE International Conference on Cloud Computing and Intelligence Systems（IEEE CCIS2016）

作者： Yifan Yuan Junping Du Jang Myung Lee Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia School of Computer ScienceBeijing University of Posts and Telecommunications Department of Electronics Engineering Pusan National University

ISBN: (纸本)9781509012572

LDA(Latent Dirichlet Allocation) model is a kind of unsupervised learning model which can extract the hidden topic from text in recent years. In this paper,we proposed a novel LDA model based on the traditional LDA model, which is integrated into the information of text category(Activity-topic LDA). In this paper, the Activity-topic LDA is proposed to improve the original latent Dirichlet allocation(LDA)model. On the basis of the LDA, the proposed method adds the tourism activity information, and obtains the probability distribution model of the tourism *** on this model, we can identify and discover the theme of tourism activities.

关键词： Latent Dirichlet allocation Text mining Gibbs sampling

来源：评论

学校读者我要写书评

暂无评论

VIDEO SUPER-RESOLUTION BASED ON NONLINEAR MAPPING AND PATCH SIMILARITY

VIDEO SUPER-RESOLUTION BASED ON NONLINEAR MAPPING AND PATCH ...

引用

2016 4th IEEE International Conference on Cloud Computing and Intelligence Systems（IEEE CCIS2016）

作者： Linghui Li Junping Du Meiyu Liang JangM yung Lee Luoming Meng Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia School of Computer ScienceBeijing University of Posts and Telecommunications Department of Electronics Engineering Pusan National University State Key Laboratory Network & Switching Technology Beijing University of Posts and Telecommunications

ISBN: (纸本)9781509012572

Aiming at recovering lost detailed information in low-resolution videos, a novel video super-resolution algorithm based on nonlinear mapping and patch similarity was proposed in this paper. Taking advantage of the strong fitting ability of the convolutional neural network for the nonlinear relationship, this algorithm utilizes the learned reconstruction parameters to pad missing structures and make the low-resolution video approach to the ground truth. However, it is unavoidable to introduce irrelevant information and amplify noises in this process. The patch similarity is imported to improve the spatial-temporal consistency and anti-noise ability of this algorithm. Moreover, instead of considering the nonlocal self-similarity only, this algorithm combines the nonlocal external sparsity and spatial-temporal similarity to enrich the prior information. All these patch similarities are used to optimize the intermediate video achieved from the mapping process. Experimental results demonstrated that the proposed algorithm achieves a competitive super-resolution quality on both the subjective and the objective evaluations, when compared to other state-of-the-art algorithms.

关键词： video super-resolution reconstruction convolutional neural network nonlocal similarity

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：