Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/191253
Title: Multimedia Mining and Representation of Content Management System for Integrated Warehouses
Researcher: Naveen Aggarwal
Guide(s): Nupur Prakash and Sanjeev Sofat
University: Guru Gobind Singh Indraprastha University
Completed Date: 2011
Abstract: The profusion of multimedia data resources over various communication networks have led revolutionary developments in the field of data mining due to different challenges posed by multimedia data complexity and variability in their representations. Multimedia mining finds its genesis in amalgamation of data mining with other computer science fields which promises to provide a cost and computation effective mechanism to represent and query the data using natural language. Multimedia Mining System helps to find interesting patterns in video by representing low level features and high level features as perceived by users, using different visual or semantic dimensions, specific to application domain. It has been observed that most of the research work in this area is very dispersed and has its own limitations. newlineIn this thesis, initial work revolves around the representation of multimedia data in different dimensions based on its content and context, such as, segmenting the videos temporally into different shots. To detect shots in video, a method has been proposed which is based on region matching using area correlation that combines different frames of video based on their statistical similarity. In this method, an area correlation energy function in consecutive frames is used to detect the similarity among the different regions of frames. The proposed method has been adopted by different researchers for temporally segmenting the medical videos used in mastoidectomy surgery. Another method is proposed to classify different production effects in video by detecting different gradual transitions such as hard cuts, fades, wipes etc. This method is tested on the benchmark videos provided by TRECVID and FISCI ILAR and demonstrates very high efficiency for news and sports video.
Pagination: 
URI: http://hdl.handle.net/10603/191253
Appears in Departments:University School of Information and Communication Technology

Files in This Item:
File Description SizeFormat 
01 tilte certificate ack abst toc.pdfAttached File4.09 MBAdobe PDFView/Open
02 chapter 1.pdf5.19 MBAdobe PDFView/Open
03 chapter 2.pdf12.06 MBAdobe PDFView/Open
04 chapter 3.pdf5.93 MBAdobe PDFView/Open
05 chapter 4.pdf6.55 MBAdobe PDFView/Open
06 chapter 5.pdf7.11 MBAdobe PDFView/Open
07 chapter 6.pdf9.04 MBAdobe PDFView/Open
08 chapter 7.pdf12.99 MBAdobe PDFView/Open


Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.

Altmetric Badge: