检索结果-内蒙古大学图书馆

作者： Shivakuamra, Palaiahnakote Lubani, Mohamed Wong, Koksheik Lu, Tong Faculty of Computer Science and Information Technology University of Malaya Kuala Lumpur Malaysia National Key Lab for Novel Software Technology Nanjing University Nanjing China

ISBN: (纸本)9781479957514

Text detection in video is a challenging problem as it is useful in several real time applications in the field of video indexing and retrieval. Unlike existing methods that generally focus on horizontal caption or graphics text, the proposed method focuses on detecting dynamic curved text in video. The method explores the characteristics of the optical flow of text, namely, constant velocity, uniform magnitude distribution and unique angle distribution, to identify text candidates with the help of k-means clustering algorithm. We propose an iterative procedure which finds the standard deviation of text candidates between the first and its successive frames, and it terminates when there is a sudden decrease in the standard deviation values. The proposed method eliminates false text candidates based on the characteristics of optical flow at component level while retaining the potential text candidates. Then, direction guided boundary growing is proposed to traverse curved text lines in video. Furthermore, the characteristics of optical flow of text are utilized at block level to eliminate false positives. Experiments are conducted with various videos, including video with static text, static and dynamic text, and dynamic text only, to evaluate the proposed method. The results are benchmarked with the existing methods to verify the superiority of our method over the existing methods in terms of recall, precision, F-measure and average processing time. © 2014 IEEE.

关键词： Optical flows

来源：评论

学校读者我要写书评

暂无评论

Multi-task rank learning for image quality assessment 40

Multi-task rank learning for image quality assessment

引用

40th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015

作者： Xu, Long Li, Jia Lin, Weisi Zhang, Yongbing Ma, Lin Fang, Yuming Zhang, Yun Yan, Yihua Key Laboratory of Solar Activity National Astronomical Observatories Chinese Academy of Sciences Beijing China State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering BeihangUniversity Beijing China International Research Institute for Multidisciplinary Science Beihang University Beijing China Department of Computer Engineering Nanyang Technological University Singapore Graduate School at Shenzhen Tsinghua University Shenzhen China Huawei Noah's Ark Lab Hong Kong School of Information Technology Jiangxi University of Finance and Economics Nanchang China Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen China

ISBN: (纸本)9781467369978

In practice, multiple types of distortions are associated with an image quality degradation process. The existing machine learning (ML) based image quality assessment (IQA) approaches generally established a unified model for all distortion types, or each model is trained independently for each distortion type by using single-task learning, which lead to the poor generalization ability of the models as applied to practical image processing. There are often the underlying cross relatedness amongst these single-task learnings in IQA, which is ignored by the previous approaches. To solve this problem, we propose a multi-task learning framework to train IQA models simultaneously across individual tasks each of which concerns one distortion type. These relatedness can be therefore exploited to improve the generalization ability of IQA models from single-task learning. In addition, pairwise image quality rank instead of image quality rating is optimized in learning task. By mapping image quality rank to image quality rating, a novel no-reference (NR) IQA approach can be derived. The experimental results confirm that the proposed Multi-task Rank Learning based IQA (MRLIQ) approach is prominent among all state-of-the-art NR-IQA approaches. © 2015 IEEE.

关键词： image quality assessment machine learning MOS pairwise comparison Rank learning

来源：评论

学校读者我要写书评

暂无评论

Detecting code reuse in android applications using component-based control flow graph

Detecting code reuse in android applications using component...

引用

作者： Sun, Xin Zhongyang, Yibing Xin, Zhi Mao, Bing Xie, Li State Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University China

ISBN: (纸本)9783642554148

Recently smartphones and mobile devices have gained incredible popularity for their vibrant feature-rich applications (or apps). Because it is easy to repackage Android apps, software plagiarism has become a serious problem. In this paper, we present an accurate and robust system DroidSim to detect code reuse. DroidSim calculates similarity score only with component-based control flow graph (CB-CFG). CB-CFG is a graph of which nodes are Android APIs and edges represent control flow precedence order in each Android component. Our system can be applied to detect repackaged apps and malware variants. We evaluate DroidSim on 121 apps and 706 malware variants. The results show that our system has no false negative and a false positive of 0.83% for repackaged apps, and a detection ratio of 96.60% for malware variants. Besides, ADAM is used to obfuscate apps and the result reveals that ADAM has no influence on our system. © IFIP International Federation for Information Processing 2014.

关键词： Android (operating system)

来源：评论

学校读者我要写书评

暂无评论

Scenery image annotation based on enhanced region semantic analysis 3

Scenery image annotation based on enhanced region semantic a...

引用

3rd International Workshop on Advanced Computational Intelligence and Intelligent Informatics, IWACIII 2013

作者： Lin, Ji Zhang, Jing Department of Computer Science and Engineering East China University of Science and Technology Shanghai China State Key Lab. for Novel Software Technology Nanjing University China

Automatic image annotation is a very essential technology cause the huge development of the multimedia and Internet. Recently, many approaches change multi-label annotation problem to single-label annotation problem while they are always time-consuming and useless to some extent. In this paper, we propose an enhanced region semantic analysis algorithm for scenery images annotation. It contains segmentation, clustering and mapping processes. We use the classical segmentation algorithm: normalized cuts and cluster patches with feature weight selection. Finally we relate cluster centers and keywords using statistical method. Experimental results show that our algorithm achieves promising performance with the scenery images and outperforms region semantic analysis algorithm on the same benchmark datasets.

关键词： Image annotation

来源：评论

学校读者我要写书评

暂无评论

Community and Social Feature-Based Multicast in Opportunistic Mobile Social Networks

Community and Social Feature-Based Multicast in Opportunisti...

引用

International Conference on computer Communications and Networks (ICCCN)

作者： Charles Shang Britney Wong Xiao Chen Wenzhong Li Suho Oh Department of Computer Science University of Illinois at Urbana-Champaign IL USA Department of Computer Science Cornell University Ithaca NY USA Department of Computer Science Texas State University San Marcos TX USA State Key Laboratory for Novel Software Technology Nanjing University China Department of Mathematics Texas State University San Marcos TX USA

Opportunistic Mobile Social Networks (OMSNs), formed by people moving around carrying mobile devices such as smartphones, PDAs, and laptops, have become popular in recent years. The OMSNs we discuss here are a special kind of delay tolerant networks (DTNs) that help enhance spontaneous interaction and communication among users that opportunistically encounter each other, without additional infrastructure support. Multicast is an important routing service in OMSNs which supports the dissemination of messages to a group of users. Most of the existing multicast algorithms are designed for general-purpose DTNs where social factors are neglected or reflected in static social features which are not updated to catch nodes' dynamic contact behavior. In this paper, we introduce the concept of dynamic social features and its enhancement to capture nodes' dynamic contact behavior, consider more social relationships among nodes, and adopt the community structure in the multicast compare-split scheme to select the best relay node for each destination in each hop to improve multicast efficiency. We propose two multicast algorithms based on these new features. The first community and social feature-based multicast algorithm is called Multi-CSDO which involves destination nodes only in community detection, and the second one is called Multi-CSDR which involves both the destination nodes and the relay candidates in community detection. The analysis of the algorithms is given and simulation results using a real trace of an OMSN show that our new algorithms outperform the existing one in terms of delivery rate, latency, and number of forwardings.

关键词： Heuristic algorithms Relays Algorithm design and analysis Routing Detection algorithms Mobile communication Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Imbalanced web spam classification based on nested rotation forest

引用

ICIC Express Letters 2015年第3期9卷 937-944页

作者： Fang, Xiaonan Zhang, Huaxiang Gao, Shuang Tan, Yanyan Department of Information Science and Engineering Shandong Normal University No. 88 Wenhua East Road Jinan China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Shandong Normal University No. 88 Wenhua East Road Jinan China

Web spam has emerged to be a critical problem in web search. Unfortunately, single classifiers always perform poorly on imbalanced web spam data sets. For better solving these problems, a nested Rotation Forest structure is proposed, i.e., ensemble classifiers instead of single decision trees are used as the base classifiers of Rotation Forest. Meanwhile, SMOTE is adopted to change the imbalanced ratio of the training set. Rotation Forest and SMOTE method are combined to improve the classification effect in the paper. Experimental results on WEBSPAM UK2007 data set show that our proposed methods can markedly boost the performance of web spam classification. © 2015, ICIC International.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

A consensus pattern of content feature and link feature for web spam detection

引用

Journal of Computational Information Systems 2014年第9期10卷 3759-3766页

作者： Gao, Shuang Zhang, Huaxiang Liu, Li Fang, Xiaonan Department of Information Science and Engineering Shandong Normal University Jinan 250014 China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Jinan 250014 China

Web spam detection is a challenging problem while there're two distinct views of web spam dataset, content view and link view. We tackle this problem by taking the different formulations and statistical properties of content and link views into full consideration. We reformulate PCA for content feature and LPP for link feature and then incorporate them into our method to drive the consensus pattern from multiple embeddings of multiple representations. We also provide a method to compute the consensus pattern for out-of-sample data points. The consensus patterns of the training data are used to construct classifiers and our experimental results demonstrate that our approach to web spam detection outperforms the related dimensionality reduction approaches. © 2014 Binary Information Press.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Improving SVM classifiers with link structure for web spam detection

引用

Journal of Computational Information Systems 2014年第6期10卷 2435-2443页

作者： Gao, Shuang Zhang, Huaxiang Zheng, Xiyuan Fang, Xiaonan Department of Information Science and Engineering Shandong Normal University Jinan 250014 China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Jinan 250014 China

Web link structure has attracted great attention in web spam detection. In this paper, we modify the SVM classifier by exploiting web link structure. We firstly construct the link structure preserving within-class scatter matrix with direct link matrix and indirect link matrix. Then we incorporate web link structure into SVM classifier to reformulate an optimization problem. The proposed method has taken advantage of the link information on the web. Experimental results on web spam datasets show that such combination of web link structure and SVM classifier can significantly outperform related methods and demonstrate the change of classification accuracy along with the round of indirect links. Copyright © 2014 Binary Information Press.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Large-scale supervised multimodal hashing with semantic correlation maximization 28

Large-scale supervised multimodal hashing with semantic corr...

引用

28th AAAI Conference on Artificial Intelligence, AAAI 2014, 26th Innovative Applications of Artificial Intelligence Conference, IAAI 2014 and the 5th Symposium on Educational Advances in Artificial Intelligence, EAAI 2014

作者： Zhang, Dongqing Li, Wu-Jun Shanghai Key Laboratory of Scalable Computing and Systems Department of Computer Science and Engineering Shanghai Jiao Tong University China National Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University China

ISBN: (纸本)9781577356790

Due to its low storage cost and fast query speed, hashing has been widely adopted for similarity search in multimedia data. In particular, more and more attentions have been payed to multimodal hashing for search in multimedia data with multiple modalities, such as images with tags. Typically, supervised information of semantic labels is also available for the data points in many real applications. Hence, many supervised multimodal hashing (SMH) methods have been proposed to utilize such semantic labels to further improve the search accuracy. However, the training time complexity of most existing SMH methods is too high, which makes them unscalable to large-scale datasets. In this paper, a novel SMH method, called semantic correlation maximization (SCM), is proposed to seamlessly integrate semantic labels into the hashing learning procedure for large-scale data modeling. Experimental results on two real-world datasets show that SCM can significantly outperform the state-of-the-art SMH methods, in terms of both accuracy and scalability. Copyright © 2014, Association for the Advancement of Artificial Intelligence.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Supervised hashing with latent factor models 14

Supervised hashing with latent factor models

引用

37th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2014

作者： Zhang, Peichao Zhang, Wei Li, Wu-Jun Guo, Minyi Shanghai Key Laboratory of Scalable Computing and Systems Department of Computer Science and Engineering Shanghai Jiao Tong University China National Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University China

ISBN: (纸本)9781450322591

Due to its low storage cost and fast query speed, hashing has been widely adopted for approximate nearest neighbor search in large-scale datasets. Traditional hashing methods try to learn the hash codes in an unsupervised way where the metric (Euclidean) structure of the training data is preserved. Very recently, supervised hashing methods, which try to preserve the semantic structure constructed from the semantic labels of the training points, have exhibited higher accuracy than unsupervised methods. In this paper, we propose a novel supervised hashing method, called latent factor hashing (LFH), to learn similarity-preserving binary codes based on latent factor models. An algorithm with convergence guarantee is proposed to learn the parameters of LFH. Furthermore, a linear-time variant with stochastic learning optimization is proposed for training LFH on large-scale datasets. Experimental results on two large datasets with semantic labels show that LFH can achieve superior accuracy than state-of-the-art methods with comparable training time. Copyright 2014 ACM.

关键词： Big data

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：