Table of Contents
An Overview of the Informedia Digital Video Library
Project
Why Video?
Informedia Digital Video Library Project
Informedia-1 Overview
Application Areas
Application of Diverse Technologies
Challenge: Imperfect Technology
Informedia Current Capabilities
Informedia Library Creation
Informedia Library Exploration
Techniques Underlying Abstractions
Image Processing
Goal: Automatic Video Characterization
Camera and Motion Detection
Text and Face Detection
Video OCR
Video OCR Block Diagram
Video Frames being Processed
VOCR Preprocessing Problems
Video OCR
Image Search & Similarity Matching
Image Similarity Challenge
Images Containing Similar Colors
Images Containing Similar Shapes
Images Containing Similar Content
Variety of Similarity
Image Understanding Results
Speech Recognition
Speech Recognition Accuracy
Speech Recognizer Results
“Name-It” Face/Name Association
Deriving “Matching Shots”
Query-based Thumbnail Menu
Query-based Thumbnail Selection Process
“Skim Video”: Extraction of Significant
Content
KNN-based Topic Detection
Annotations
Geocoding Procedure, Part 1 of 2
Geocoding Procedure, Part 2
of 2
Named Entity Extraction
Informedia Focus
Key Challenges for Informedia-2
PPT Slide
PPT Slide
PPT Slide
PPT Slide
PPT Slide
Accomplishments
Credits |