A DISTRIBUTED MULTIMEDIA DATA MANAGEMENT OVER THE GRID KASTURI

89590 DISTRIBUTED SYSTEMS WHY DISTRIBUTED SYSTEMS?
!doctype Html html Langenus head meta Charsetutf8script Typetextjavascript(windownreum||(nreum{}))init{privacy{cookiesenabledtrue}ajax{denylist[bamnrdatanet]}distributedtracing{enabledtrue}}(windownreum||(nreum{}))loaderconfig{}! for
1 CS451 INTRODUCTION TO PARALLEL AND DISTRIBUTED COMPUTING

1 FM 92XII GRIB GENERAL REGULARLYDISTRIBUTED INFORMATION IN
13 DISTRIBUTED BY VERITAS VERITAS MAKES EVERY EFFORT TO
15 DISTRIBUTED BY VERITAS VERITAS MAKES EVERY EFFORT TO

Distributed computing uses the computing power of multiple computers or clusters of computers to provide the user with a virtual supercomputer

A Distributed Multimedia Data Management over the Grid


Kasturi Chatterjee


Distributed computing uses the computing power of multiple computers or clusters of computers to provide the user with a virtual supercomputer. Grid computing is a form of distributed computing which combines the power of several computers of varied computing resources to execute one or more task collaboratively in a seamless and transparent manner. In 2000, Ian Foster along with Steve Tuecke, defined that Grid computing is concerned with “coordinated resource sharing and problem solving in dynamic, multi-institutional virtual organizations” [1]. Many applications leverage the huge computing power of Grid to make things faster and efficient. Multimedia applications are one of the most popular and most used applications in today’s world. With the proliferation of internet and information technology, the availability of numerous sources of video, images, text and other multimedia data has become easy. They are used for varied purposes like education, entertainment, business etc. The sheer size of multimedia data like a video makes distributing it over a shared and distributed environment and utilizing the computing power of grids a likely choice. A distributed environment storing multimedia data will enable a user to share and use the multimedia data from a distant node without wasting storage space in his/her own local machine by making a duplicate copy. Thus, a connected group of users forming a grid can utilize the collective computing power and storage of multiple computers to solve different issues. But the atypical nature of multimedia data makes its management, especially in a distributed environment, a daunting task. The two most important characteristics of multimedia data that makes it different from traditional data are (1) high dimensionality and (2) semantic interpretation. Multimedia data is represented as multidimensional feature vectors and semantic meanings are tagged to them by the users based on perception. These two characteristics of multimedia data provides the impetus for developing special techniques to better manage them. Here, we propose an efficient multimedia data management over a distributed environment like grid.

Management of multimedia data involves three important parts: Firstly, the extraction of content /features from the videos/images. This is a very important step since to identify the feature sets that best describes a multimedia content determines the effectiveness and success of the subsequent processing. Secondly, storing and indexing of multimedia data. The huge volume of multimedia data requires an efficient storage and indexing which takes care of the special nuances of multimedia data viz. the high dimensionality and the gap between low-level features and high-level semantic interpretation. Finally, efficient retrieval of multimedia data. The most popular among the different retrieval approaches is content-based image/video retrieval. In this project, we propose to primarily achieve the second part, which is the backbone of efficient multimedia data management, over a grid. However, to achieve it efficiently, effective execution of the first and third parts is equally crucial. We will implement a multidimensional index structure, the affinity hybrid tree [2], in a distributed environment and devise content-based retrieval algorithms supported by the index structure over a grid. The index structure and the subsequent retrieval methodologies coupled with it, is independent of the features used to represent the multimedia data and the techniques adopted to interpret users’ preference during content-based retrieval. The index structure demonstrates encouraging results in single node architecture both in terms of low computational overhead and relevance of query results (which measures how close the retrieval results are to the users’ perception). In the single node architecture, we used a matrix, called affinity relationship [3], which captures the high-level image/video relationship using a probabilistic model called hidden markov model mediator. Here we plan to use a different representation of the affinity relationship in the form of a hierarchical distributed and often replicated structure, to capture and use the high-level semantic relationships and utilize them while content-based retrieval effectively.



Reference:

[1] Ian Foster, “What is the Grid? A Three Point Checklist”,2002.

[2] Kasturi Chatterjee and Shu-Ching Chen, "A Novel Indexing and Access Mechanism using Affinity Hybrid Tree for Content-Based Image Retrieval in Multimedia Databases," International Journal of Semantic Computing (IJSC), Vol. 1, Issue 2, pp. 147-170, June 2007.

[3] Mei-Ling Shyu, Shu-Ching Chen, Min Chen, Chengcui Zhang, and Chi-Min Shu, "MMM: A Stochastic Mechanism for Image Database Queries," Proceedings of the IEEE Fifth International Symposium on Multimedia Software Engineering (MSE2003), pp. 188-195, December 10-12, 2003, Taichung, Taiwan, ROC.



30 GUIDE TO NOTES OVERVIEW MULTISTEP AND DISTRIBUTED PROCESSING
39 DISTRIBUTED BY VERITAS VERITAS MAKES EVERY EFFORT TO
6 DISTRIBUTED BY VERITAS TRUST TEL [263] [4] 794478


Tags: distributed multimedia, hierarchical distributed, multimedia, management, distributed, kasturi