What is this?
Agenda
Meetings
Email Archive
   - Autumn
   - Spring
   - Summer
Misc
Documentation
Useful Links

Download!

Image Indexing and Retrieval: A Content-based Approach (gp-qiu1)

The rapid expansions of the Internet and fast advance in digital imaging technology have enabled the wide spread use of image libraries and databases. One of the challenges in managing large image databases is the indexing of images. There are generally two approaches to indexing textual-based and content-based approaches. In textual based approach, an image is annotated with a set of keywords that appropriately describe the visual content of the image. Database access/image retrieval in this approach is based on keyword (text) search/matching. In contrast, content-based approach uses advanced image processing/computer vision techniques to extract a set of visual features (such as colour distributions, texture characteristics, objects shapes etc) from the image as its annotation. In this second approach, database access/image retrieval is based on "query by example" whereby the user presents the system with an example (image) and retrieves from the database those images that are similar (to the example).

Last year, a group of student has successfully developed an image indexing and retrieval system based on text. Your task is to extend the system to enable image indexing and retrieval based on visual contents.