Please use this identifier to cite or link to this item:
Title: Classification of MPEG videos based on motion vectors
Authors: Zaw Tin Oo
Keywords: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
Issue Date: 1999
Abstract: This dissertation provides a method of classifying MPEG videos based on Motion Vector. We developed a prototype system that can extract the motion vectors embedded in the video so as to identify the region of interest for segmentation and classification purpose. The two classifications are identified as wide angle shot and neighboring shot. We tested various MPEG-1 and MPEG-2 standard videos and classify them as wide angle and neighboring shots. By extracting the accurate motion characteristics of the video, we are able to classify video sequences based on their different characteristics. We developed tools based on camera motion for analyzing and classifying a class of structured videos using the motion information available directly from MPEG compressed videos. By examining readily extractable data from MPEG video bit streams, such as macroblock motion vectors and the number of intra-coded blocks, we illustrate that it is possible to identify high level events such as wide angle and neighboring views.
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
  Restricted Access
Main report10.74 MBAdobe PDFView/Open

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.