| Description |
Emerging video-enabled products, such as mobile phones or portable TV devices, facilitate the use of video surveillance systems. An essential task in video surveillance is the classification of video objects into different classes such as "people and vehicles" or "cars and vans", or into "sedan and taxi".
In this project, students will be guided to first implement a video system that classifies objects into their respected classes. To train the implemented system, the students will develop a system to label a database of surveillance video signals, where each video object will be assigned one of the classes in the classification system.
The team will explore an existing graphical tool for labeling of video databases and extend that tool so that it will be suitable to label video objects for object classification tasks semi-automatically.
Finally, the team will use XML (Extensible Markup Language) to store class and object information. They will need to write an XML parser to read class and object information. XML facilitates cross-platform storage and exchange of information in a user-friendly, extendable and efficient manner. XML is used by many public video datasets to provide class labeling information.
Student Requirements: Student should have good knowledge in Matlab, C++, and GUI design.
|