Please use this identifier to cite or link to this item: http://hdl.handle.net/123456789/14411
Title: Temporal Video Scene Segmentation
Authors: Kumar, Manoj
Keywords: Temporal Video Scene Segmentation
Computer Vision Problem
Shot Similarity Graph (SSG)
Sliding Window Method
Issue Date: May-2016
Publisher: Computer Science and Engineering,IITR.
Abstract: This Dissertation report discusses one of the computer vision problem, which is implemented using various video and image processing techniques from last two decades that is “TEMPORAL VIDEO SCENE SEGMENTATION”. From last decade a lot of work has been done on automatically video scene segmentation. In field of computer vision this problem has received most popularity because it is the first and most important part of the other problems like video summarization, video indexing and browsing etc. This dissertation includes a new approaches to solve this problem which consists of different feature detector and descriptors, clustering algorithms, window based scene boundary defining etc. As we know scenes are the grouping of semantically similar shots which are temporarily close. So our first task is divide the video into shots, which includes cluster of similar frames. This task is done using HSV color histogram. Now we extract some visual, motion and SIFT features of the shots and calculated inter shot pairs similarity, which is interpreted with shot similarity graph (SSG). At the final stage using sliding window method we have grouped similar shots which are higher similarity then some decided threshold. This grouping has considered inverse time proximity. This dissertation has also include some of implemented papers with their different approaches and achieved results are also discussed.
URI: http://hdl.handle.net/123456789/14411
Appears in Collections:DOCTORAL THESES (E & C)

Files in This Item:
File Description SizeFormat 
G25973-MANOJ-D.pdf1.81 MBAdobe PDFView/Open


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