Transcript Why Digital Video Library?

Presentation Outline
Project Aims
Introduction of Digital Video Library
Introduction of Our Work
Considerations and Approach
Design and Implementation
Future Plans
Project Aims
Learn different issues about DVL
existing such as Informedia(TM) done by CMU
Implement a small-scale model
Client interface for playback of media
Process data source/user queries in Chinese
Web Integration
Introduction of Digital Video Library
Functions of a DVL
Why Digital Video Library?
Issues about DVL
Techniques Addressing DVL Issues
Functions of a DVL
Storage and play back of video resources
Online access of video resources
Search of useful information through the vast
data collections
Retrieval of the most relevant selections
Why Digital Video Library?
Digital data
easy to process in an automated way
takes advantage of the Internet
Multimedia content
more attractive
for education, training, sports and entertainment
Library management
An organized way of retrieval media contents
Issues about DVL
Building Video Databases
Indexing the Video Contents
Breaking the Video into Segments
Retrieving Video
Building Video Database
Extremely large space requirement
Choosing a video format
Online access of video resources
Indexing the Video Contents
Inaccessibility of information in video
Limitation of text-base descriptions
Tedious manual process
Cinematic information is complex to describe
Breaking the Video into Segments
Video paragraphing
Alternate representation of video
Text tile
Text abstract
Full-text transcript
Representative single image
“Skim” video
Retrieving Video
Three principle for searching
Let user to skim the video object quickly with their interest
Adjust the size of video
Identify of desired video among multiple returned results
Ways to achieve
Small size of video returned
Information visualization
Overview of
Digital Video Library System
Techniques Addressing DVL Issues
Speech Analysis
Speech recognition technology generates transcript
Pauses and silence in audio help paragraphing
Image Analysis
Detect significant changes in weighted color histogram
Interpret camera motion
Natural Language Processing
Query processing
Retrieval
Our Work
Focus and Target
Programming Environment
Design and Implementation
System Consideration
System Design
System Implementation
Future Plans
Focus and Target
Divide the project into modules that can be
implemented in stages
digital video preparation
video playback
user query
web-integration
Programming Environment
Platform
Java as programming language
JMF API to present time-base media
Why use Java?
Platform independence
Network ready
International appeal
Easy to build GUI
JMF API for video playback
Disadvantages?
Presenting Time-Base Media with
JMF
What is JMF?
Java Media Framework Application Programming Interface
Incorporating media data into Java
The high level architecture
Similar to traditional VCR
Data Source
Pull data-source and Push data-source
Media Presentation
Programmed or Interactive controls
Design and Implementation
System Consideration
System Design
System Implementation
System Consideration
User Perspective
User friendly
Other system component are transparent
Server Program
Robust, reliable
Network Issue
Video Collections
System Design
Client Program
Query request
Query request
Serv er respond
Watch the v ideo
User
Video Retriv al
Query Server
searching in database
Library Database
System Implementation
Library Preparation
Server Program
Client Program
Future Plans
Java-applet Client Program
Semi-auto Timestamp Editing Tools
Searching and Indexing Capability
Other Possible Extensions
Automatically derived transcripts
Language processing of queries and transcripts
Content-based Image Manipulation