检索结果-内蒙古大学图书馆

作者： Sanaa Chafik Telecom SudParis

学位级别：博士

The amount of media data is growing at Our contributions can be classiﬁed into shallow and high speed with the fast growth of Internet and media deep approaches. In the fonner category, we propose resources. Performing an efficient similarity (nearest two shallow hashing-based approaches, namely, neighbor) search in suchalarge collection of data isa Symmetries of the Cube Locality Sensitive Hashing very challenging problem that the scientiﬁc (SC-LSH), and Cluster-based Data Oriented Hashing community has been attempting to tackle. One of the (CDOH), based respectively on randorr1ized-hashing most promising solutions to this fundamental and shallow learning-to-hash schemes. The SC-LSH problem is Content-Based Media retrieval (CBMR) method provides a solution to the space storage systems. The latter are search systems that perform problem faced by most randomized-based hashing the retrieval task in large media databases based on approaches. It consists of a semi-random scheme the content of the data. CBMR systems consist reducing partially the randomness effect, and thus the essentially of three major units, a Data memory storage problem, of these approaches, while Representation unit for feature representation, a maintaining their efficiency in structuring Multidimensional Indexing unit for structuring the heterogeneous spaces. The CDOH approach proposes resulting feature space, and a Nearest Neighbor to eliminate the randomness effect by combining Search unit to perform efficient search. Media data unsupervised machine learning techniques with the (i. e. image, text, audio, video, etc.) can be hashing concept. CDOH outperforms the randomized represented by meaningful numeric infom1ation(i. e. hashing approaches in terms of computation time, multidimensional vector), called Feature, memory space and search accuracy. describing the overall content of the input data. The The third contribution is a deep leaming-based hashing task of the second unit is to structure the

关键词： Information retrieval Content-Based Information retrieval Multidimensional ing Hashing Nearest Neighbor Search Deep Learning Unsupervised Learning

来源：评论

学校读者我要写书评

暂无评论

Web-based Multimedia Research and Indexation for Big Data databases 3

Web-based Multimedia Research and Indexation for Big Data Da...

引用

3rd International Conference of Cloud Computing Technologies and Applications (CloudTech)

作者： Belarbi, Mohammed Amin Mahmoudi, Said Belalem, Ghalem Mahmoudi, Sidi Ahmed Univ Mons Fac Engn 20 Pl Parc Mons Belgium Ahmed Ben Bella Univ Oran1 Fac Exact & Appl Sci Oran Algeria

ISBN: (纸本)9781538611159

Indexing images represents a necessary tool for various domains related to computer vision such as video surveillance and movements analysis. Visual characteristics of image such as color, texture, shape are used to identify the content of these frames. The high increase of data (Big Data) makes us face a problem of curse of dimensionality. In this paper, we propose a cloud-based technique for images indexing images using PCA (Principal component analysis) and a binary tree. Our work consists of exploiting SIFT (Scale Invariant Feature Transform) and SURF (Speeded up Robust Features) descriptors as frames features, PCA as dimensionality reduction method of SIFT and SURF features and the binary tree representation of these images in order to accelerate the research phase and reduce the storage space. Our approach consists of developing a system based on three phases: indexing images with SIFT and SURF features, compression within the application of PCA and finally the image retrieval in the cloud.

关键词： Indexing images SIFT SURF PCA Cloud Big Data

来源：评论

学校读者我要写书评

暂无评论

Large-scale video copy retrieval with temporal-concentration SIFT

引用

NEUROCOMPUTING 2016年 187卷 83-91页

作者： Zhu, Yingying Huang, Xiaoyan Huang, Qiang Tian, Qi Shenzhen Univ Sch Comp Sci & Software Engn Shenzhen 518060 Peoples R China Shenzhen Co Ltd Oracle Res & Dev Ctr Shenzhen 518057 Peoples R China Univ Texas San Antonio Dept Comp Sci San Antonio TX 78249 USA

The scale-invariant feature transform (SIFT) feature plays a very important role in multimedia content analysis, such as near-duplicate image and video retrieval. However, the storage and query costs of SIFT become unbearable for large-scale databases. In this paper, SIFT features are robustly encoded with temporal information by tracking the SIFT to generate temporal-concentration SIFT (TCSIFT), which highly compresses the quantity of local features to reduce visual redundancy, and keeps the advantages of SIFT as much as possible at the same time. On the basis of TCSIFT, a novel framework for large-scale video copy retrieval is proposed in which the processes of retrieval and validation are implemented at the feature and frame level. Experimental results for two different datasets, i.e., CC_WEB_video and TRECVID, demonstrate that our method can yield comparable accuracy, compact storage size, and more efficient execution time, as well as adapt to various video transformations. (C) 2015 Elsevier B.V. All rights reserved.

关键词： video copy retrieval SIFT Spatio-temporal features Frame validation

来源：评论

学校读者我要写书评

暂无评论

Efficient image retrieval Using image and Audio Features in video Stream

Efficient Image Retrieval Using Image and Audio Features in ...

引用

International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS)

作者： In-Kyoung Shin Hyochang Ahn Yong-Hwan Lee Dept. of Applied Computer Engineering Dankook University Yongin-si Korea Dept. of Smart Mobile Far East Uinversity Chungcheongbuk-do Korea

ISBN: (纸本)9781509009855

Recently, information has changed to the development of the information communication network and multimedia technology. The multimedia technology is used as application techniques including the storage, playback, compression, and so on. Also it has changed rapidly and extensively over the world. Therefore numerous search schemes such as the video retrieval, speech retrieval are required. Current video retrieval methods have various problems as following redundancy of data and the inconsistency of a keyword, and infringement of copyright. In this paper, we propose implementation of novel scheme that carried out efficient retrievals for audio signal and image in video stream as rapidly changing generation of information communication.

关键词： Feature extraction Streaming media Motion pictures Speech Speech recognition databases Multimedia communication

来源：评论

学校读者我要写书评

暂无评论

视频局部特征描述子的紧凑表示方法

引用

中国图象图形学报 2016年第3期21卷 390-395页

作者：张翔王诗淇张新峰马思伟高文北京大学信息科学与技术学院数字媒体研究所北京100871

目的随着手持移动设备的迅猛发展和大数据时代的到来,以多媒体数据为核心的视觉搜索等研究和应用得到了广泛关注。其中局部特征描述子的压缩、存储和传输起到了举足轻重的作用。为此在传统图像/视频压缩框架中,提出一种高效的视觉局部... 详细信息

目的随着手持移动设备的迅猛发展和大数据时代的到来,以多媒体数据为核心的视觉搜索等研究和应用得到了广泛关注。其中局部特征描述子的压缩、存储和传输起到了举足轻重的作用。为此在传统图像/视频压缩框架中,提出一种高效的视觉局部特征的紧凑表示方法,使得传统内容编码可以适应广泛的检索分析等需求。方法为了得到紧凑、有区分度、同时高效的局部特征表示,首先引入了多参考的预测机制,在消除了时空冗余的同时,通过充分利用视频纹理编码的信息,消除了来自纹理-特征之间的冗余。此外,还提出了一种新的率失真优化方法——码率-准确率最优化方法,使得基于匹配/检索应用的性能达到最优。结果在不同数据集上进行验证实验,和最新的视频局部描述子压缩框架进行比较,本文方法能够在保证匹配和检索性能的基础上,显著地减少特征带来的比特消耗,达到大约150∶1的压缩比。结论本文方法适用于传统图像/视频编码框架,通过在码流中嵌入少量表示特征的信息,即可实现高效的检索性能,是一种面向检索等智能设备应用的新型多媒体内容编码框架。

关键词：计算机科学与技术视觉搜索视频局部特征描述子尺度不变特征变换高效视频压缩

来源：评论

学校读者我要写书评

暂无评论

An efficient content-based video retrieval for large databases

An efficient content-based video retrieval for large databas...

引用

IEEE International Conference on Mechatronics, Electronics and Automotive Engineering (ICMEAE)

作者： Cedillo-Hernandez, Antonio Cedillo-Hernandez, Manuel Garcia-Ugalde, Francisco Nakano-Miyatake, Mariko Perez-Meana, Hector Univ Nacl Autonoma Mexico Fac Ingn Div Ingn Elect Mexico City 04510 DF Mexico Inst Politecn Nacl Escuela Super Ingn Mecan & Elect Mexico City DF Mexico

ISBN: (纸本)9781467383295

Multimedia data grows fast due to advances in information technologies, creating the demand for efficient video indexing and object retrieval techniques. Traditional methods consume significant computational resources such as storage space and processing time. In this paper we propose an efficient content-based video retrieval system that is based on three main stages. The first stage involves computing the DC-image of each I-frame, from which a summarization process to extract key-frames is performed. During the second stage, a segmentation processes is applied to each key-frame in order to isolate the region of interest within it. Local features are extracted from the resulting area and are stored as the descriptor of the frame. The retrieval stage is carried out by computing the Euclidean distance and determines if its content is related with the video database. Experimental results show that the proposed approach is promising in terms of efficiency and effectiveness.

关键词： Efficient CBVR system video indexing DC-frame SURF

来源：评论

学校读者我要写书评

暂无评论

A fast wavelet packet and PCA based image indexing and authentication method 28

A fast wavelet packet and PCA based image indexing and authe...

引用

2015 28th IEEE Canadian Conference on Electrical and Computer Engineering, CCECE 2015

作者： Alzahir, Saif Singh, D. Department of Computer Science UNBC BCV2N4Z9 Canada Computer Science Department UNBC Canada

ISBN: (纸本)9781479958276

The rapid growth in image databases is presenting new research challenges in wide range of areas that span image indexing, image retrieval, authentication, and recognition. This paper presents a new algorithm for image indexing and authentication. Although the combination of wavelet transform and principal components analysis has been used in some image processing areas but they have not been used for image indexing and authentication before. The proposed algorithm is based on the extraction of selected features of images using principal components analysis and wavelet packets. The proposed algorithm addresses three main concerns: (i) search accuracy;(ii) time efficiency;and (iii) storage requirements. We use wavelet packets to decompose the image to its fundamental subbands then we apply PCA method to extract the features of the images. The extracted features are used to create image signatures. We have tested our algorithm on the ORL Database of Faces (the AT&T laboratories Cambridge Database), PIE database, Caltech-UCSD Birds 200 and other miscellaneous images from different sources. We compared our results with two other methods [2] and [13] and found that our method retrieved the query image with 100% efficiency which significantly higher than that of method [13] and it is much faster than both methods. © 2015 IEEE.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of Distributed video retrieval System in Air Monitoring Warning System

Design and Implementation of Distributed Video Retrieval Sys...

引用

International Conference on Measuring Technology and Mechatronics Automation, ICMTMA

作者： Chen Na Zhu Ya Ling College of Software Engineering Lanzhou Institute of Technology Lanzhou China

ISBN: (纸本)9781467371445

The video monitoring system plays a decisive role on the security of the sensitive areas, the video will be recording when the system monitors a warning behavior, accumulating over a long period, which will produce a large amount of video information, It is becoming crucial that how to retrieve these massive video information effectively and efficiently. The Content-based video retrieval technology has been the research hotspots in recent years. This paper, with the mature technology of video key frame extraction, combining with the SIFT feature extraction and matching, as well as LSH mapping, realizes the video retrieval. During the process of feature extraction and matching to the video key frame, we realize the distributed processing and storage, using MapReduce parallel mode and HBase distributed storage system, which is the application development of the Content-based video retrieval on the parallel analysis and scalable storage.

关键词： Monitoring Feature extraction Indexes Distributed databases Streaming media image retrieval Distributed processing

来源：评论

学校读者我要写书评

暂无评论

An Efficient Content-Based video retrieval for Large databases

An Efficient Content-Based Video Retrieval for Large Databas...

引用

International Conference on Mechatronics, Electronics and Automotive Engineering (ICMEAE)

作者： Antonio Cedillo-Hernández Manuel Cedillo-Hernández Francisco García-Ugalde Mariko Nakano-Miyatake Héctor Pérez-Meana Facultad de Ingeniería División de Ingeniería Eléctrica Universidad Nacional Autónoma de México D.F. México Sección de Estudios de Posgrado e Investigación Instituto Politécnico Nacional D.F México

Multimedia data grows fast due to advances in information technologies, creating the demand for efficient video indexing and object retrieval techniques. Traditional methods consume significant computational resources such as storage space and processing time. In this paper we propose an efficient content-based video retrieval system that is based on three main stages. The first stage involves computing the DCimage of each I-frame, from which a summarization process to extract key-frames is performed. During the second stage, a segmentation processes is applied to each key-frame in order to isolate the region of interest within it. Local features are extracted from the resulting area and are stored as the descriptor of the frame. The retrieval stage is carried out by computing the Euclidean distance and determines if its content is related with the video database. Experimental results show that the proposed approach is promising in terms of efficiency and effectiveness.

关键词： Feature extraction video sequences image color analysis image segmentation Streaming media Indexing

来源：评论

学校读者我要写书评

暂无评论

Gestão de imagens em movimento de matérias jornalísticas em banco de dados na TV Digital

Gestão de imagens em movimento de matérias jornalísticas ...

引用

作者： Gonçalves, Maria Valéria Bertachini do Nascimento Universidade Estadual Paulista

Due to the intense information flow, together with the techonological trends of control, storage and distribution, information management is considered as a key factor enabling dynamism and skills to manage and direct this raw material that is information. In this scenario, is the television media that uses mainly the image as information while determining language for the audiovisual sector. Television journalism the image appears to portray and cover your news. The study object of this thesis is the management of movement of news stories in images, considering that this information will reach its goal if it is processed and managed in the database, following a following and promoting the distribution and use of this information its users to quickly and accurately. The central question that guided the studies and research refers to the treatment of the moving image and how it should be guided so that there is accuracy in recovery. Thus, if the objective was, in general, the optimization in activities involving information management movement of news stories in pictures in a regional television station offer theoretical and practical contribution to provide management practices for this area Digital Television. The required specific objectives resulted in the construction of a theoretical framework to support that following culminated in conducting exploratory study by the Documentation Centre - CEDOC, the broadcaster TV TEM, affiliated to Rede Globo and brinding together the cities of Sorocaba, Bauru, São Jose do Rio Preto and Itapetininga, to present guiding parameters to the moving image management conduct with the databases in Digital TV ambience. The methodological procedures to achieve the established objectives were operationalized through the development of exploratory research, with the steps of the research/literature review; field research that split into information retrieval, enabling the characterization of TEM TV and CEDOC...Devido ao fluxo informacio

关键词： videoteipes - Administração Jornalismo Televisão digital Arquivos audiovisuais Materiais audiovisuais video tapes - Management Dissertação de mestrado

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：