Project Title

An Automatic Multi-Layer Video Content Classification Framework

Project Objectives

Content-based classification of video stories enables users to find their desired stories faster and more accurately from vast amount of available video sources. Traditional video classification methods usually use texts, including captions and speech scripts, as the only clue to achieve the purpose. This project intends to classify large-scale video stories effectively by using both syntactic features and semantic features in a unified learning framework.

This project intends to advance the state-of-the-art of digital video library research and machine learning research in several aspects. The major objectives in this project can be summarized as follows:

‧ Syntactic feature and semantic feature definition criteria and extraction algorithms
‧ Visual feature extraction for images and videos
‧ Multi-layer video content class hierarchy mechanisms
‧ Unified learning framework for large-scale multimedia classification
‧ Machine learning techniques for video content classification and retrieval
　

Project Sponsor

This project is supported by grants from Shun Hing Institute of Advanced Engineering (SHIAE).