This project is to generate meaningful labels for web images by utilizing associated information to facilitate web search of visual information.
The Web has become a wealthy repository of various types of information for many tasks such as knowledge discovery. Images have been widely utilized to enhance web contents. This project is to leverage web information for image annotation, since web page contents consist of multiple types of data (e.g. text, image, and video) which are typically related to each other. Rather than empirically extract surrounding text to represent image contents, this project is to investigate novel techniques to explore the correlation between image contents and web page contents, as well as to discover new knowledge. This project will utilize our established web image repository.
The opportunity ID for this research opportunity is 318