Classifying Video Clips

Research Paper # 57640

Computer and Technology > Programming
Computer and Technology > Software
Film

A discussion and investigation of the process of classification of video clips using image processing techniques.

4,874 words () | 6 sources | MLA | 2003 United Kingdom

Paper Summary:

This paper investigates the task of automatic recognition of film genres, more specifically the recognition and classification of theatrical movie trailers on the basis of genre.

Outline
Introduction
Background
Specification and Design
Shot detection
Feature Extraction
Classification
Implementation
Reading in a Movie File
Shot Boundaries Detection and Key Frame Extraction
Key Frame Feature Extraction Class Identification
Results and Evaluation Key Frames
Intensities and Standard Deviation
Overall Project Evaluation
Future Work
Conclusion
Appendices

From the Paper:

"The use of key frame analysis will greatly reduce the computational time spent by the system obtaining feature statistics that can be used for classification. Although the computational effort is reduced the accuracy should not be that adversely affected since using one frame from each shot in the trailer will give a fairly rounded view of the content of the clip as a whole. Throughout the project every effort will be made to reduce the computational complexity of the system since Matlab is a very resource intensive environment, and a reasonable execution time is desired."