By Milan Petkovic, Willem Jonker

The region of content-based video retrieval is a really sizzling zone either for learn and for advertisement purposes. which will layout powerful video databases for purposes similar to electronic libraries, video creation, and various net functions, there's a nice have to advance powerful ideas for content-based video retrieval. one of many major concerns during this region of analysis is how you can bridge the semantic hole among low-Ievel positive aspects extracted from a video (such as colour, texture, form, movement, and others) and semantics that describe video idea on a better point. during this ebook, Dr. Milan Petkovi6 and Prof. Dr. Willem Jonker have addressed this factor via constructing and describing numerous leading edge strategies to bridge the semantic hole. the most contribution in their examine, that's the middle of the ebook, is the advance of 3 ideas for bridging the semantic hole: (1) a method that makes use of the spatio-temporal extension of the Cobra framework, (2) a method in line with hidden Markov versions, and (3) a method in accordance with Bayesian trust networks. to guage functionality of those suggestions, the authors have performed a few experiments utilizing actual video info. The publication additionally discusses domain names suggestions as opposed to normal resolution of the matter. Petkovi6 and Jonker proposed an answer that permits a method to be utilized in a number of domain names with minimum alterations. in addition they designed and defined a prototype video database administration process, that's according to thoughts they proposed within the book.

Show description

Read Online or Download Content-Based Video Retrieval: A Database Perspective PDF

Similar imaging systems books

Optics

Optics as an issue has advanced dramatically lately, with many purposes all through technological know-how and expertise.

Time-Varying Image Processing and Moving Object Recognition, 4. Proceedings of the 5th International Workshop Florence, Italy, September 5–6, 1996

. .. Researchers in snapshot conversation, man made intelligence and robotics will locate a lot of curiosity. ASLIB e-book advisor, 1997

Human Engineering in Stereoscopic Viewing Devices

This booklet gathers jointly info in regards to the interplay of hu­ guy stereopsis with a number of stereoscopic viewing units, in particular these utilized in teleoperator platforms. The e-book isn't serious about desktop vi­ sion structures. In those platforms, information analogous to human binocular visible details is accrued and analyzed via a few machine to be used in choice making or keep an eye on, usually with no the intervention of a human.

Renal Cell Carcinoma: Molecular Targets and Clinical Applications

Within the moment variation in their severely acclaimed e-book, Ronald Bukowski, Robert Motzer, and Robert Figlin have completely up-to-date and improved their survey of scientific, organic and pathological administration of localized and complicated renal phone carcinoma. A panel of across the world well known participants explores the newest advancements in molecular genetics, concentrating on the radical objectives which were found in epithelial renal tumors.

Extra info for Content-Based Video Retrieval: A Database Perspective

Sample text

Chang, W. Chen, H. Meng, H. Sundaram, D. Zhong, "A Fully Automated Content Based Video Search Engine Supporting Spatio-Temporal Queries", IEEE Transaction on Circuits and Systemsfor Video Technology, Vol. 8, No. , 1998. [33] T. T. Kuo, A. P. Chen, "A Content-Based Query Language for Video Databases", Proc. ofIEEE Multimedia Computing Systems, 1996. [34] G. Amato, G. Mainetto, P. Savino, "An Approach to a Content-Based Retrieval of Multimedia Data" Multimedia Tools and applications, Vol. 7, No. 1/2, 1998, pp.

These approaches have been strongly criticized because of their inflexibility and many limitations. Fixed video segmentation enforces the only one segmentation of the original data. A segmented part of a video is separated 41 Video Modeling from its context. Hence, the important contextual information about a video sequence is lost. On the other hand, approaches that are more flexible use semantic criteria to segment a video into logical units based on extracted content. Practically, a video is expressed with a set of descriptions that have frame addresses.

3 Searching video data As video documents involve huge amounts of data from various media, the need to efficiently query this data becomes significant. Bearing in mind 24 Content-Based Video Retrieval: A Database Perspective the nature of video data, we concentrate on specific issues of the query process in this subsection. Video documents are often weIl structured in a similar way as books are organized into chapters and sections (the next ehapter explains this in more details). Therefore, in most number of cases, the video strueture ean be queried using well-developed techniques of the existing technology.

Download PDF sample

Rated 4.38 of 5 – based on 45 votes