Abstract

The rapid advances in electronic imaging, storage, data compression telecommunications, and networking technology have resulted in a vast creation and use of digital videos in many important applications such as digital libraries, distance learning, public information systems, electronic commerce, movie on demand, etc. This brings about the need for management as well as delivery of video data. Organizing and managing video data, however, is much more complex than managing conventional text data due to their semantically rich and unstructured contents. Also, the enormous size of video files requires high communication bandwidth for data delivery. In this dissertation, I present the following techniques for video data management and delivery.

  • Decomposing video into meaningful pieces (i.e., shots) is a very fundamental step to handling the complicated contents of video data. Content-based video parsing techniques are presented and analyzed.
  • In order to reduce the computation cost substantially, a non-sequential approach to shot boundary detection is investigated.
  • Efficient browsing and indexing of video data are essential for video data management. Non-linear browsing and cost-effective indexing schemes for video data based on their contents are described and evaluated.
  • In order to satisfy various user requests, delivering long videos through the limited capacity of bandwidth is challenging work. To reduce the demand on this bandwidth, a hybrid of two effective approaches, periodic broadcast and scheduled multicast, is discussed and simulated.

The current techniques related to the above works are discussed thoroughly to explain their advantages and disadvantages, and to make the new improved schemes. The substantial amount of experiments and simulations as well as the concepts are provided to compare the introduced techniques with the other existing ones. The results indicate that they outperform recent techniques by a significant margin. I conclude the dissertation with a discussing of future research directions.

Graduation Date

2000

Semester

Fall

Advisor

Hua, Kien

Degree

Doctor of Philosophy (Ph.D.)

College

College of Engineering

Department

Electrical Engineering and Computer Science

Degree Program

Electrical Engineering and Computer Science

Format

PDF

Language

English

Rights

Written permission granted by copyright holder to the University of Central Florida Libraries to digitize and distribute for nonprofit, educational purposes.

Length of Campus-only Access

None

Access Status

Doctoral Dissertation (Open Access)

Identifier

DP0001719

Accessibility Status

Searchable text

Share

COinS