| | | |

Computational Video

Fast Image Retrieval

The aim of Fast Image Retrieval is to create technology to allow retrieval of images via something similar to a search engine. This means creating a system in which people can select an image, or a part of an image, and do a fast search to find all similar images in a large database.

The project started in July 2005 and is expected to end in July 2008.

While the web can be searched using text keywords, effective image searching does not yet exist. The critical issue is: what features should be extracted from each image? Recently, features called “interest points” that characterize distinctive parts of an image have been identified. Our system has a simple way to create an index file for interest points, which greatly speeds up image search. An off-line process creates an index file of a large number of images. The resulting interactive, online system allows a user to select an image, or a part of an image, and to quickly search large image databases for similar images.

Fast Image Retrieval has the potential to be useful in security applications and in detection of copyright violations.

Research Contact

Dr. Gerhard Roth
NRC Institute for Information Technology
1200 Montreal Road
Building M-50, Room C-327
Ottawa, ON K1A 0R6
Telephone: +1 (613) 993-1219
Fax: +1 (613) 952-0215
E-mail: Gerhard.Roth@nrc-cnrc.gc.ca

Business Contact

Dr. George Forester
Group Coordinator, Business Development
Business Development Office
NRC Institute for Information Technology
1200 Montreal Road
Building M-50, Room 203
Ottawa, ON K1A 0R6
Telephone: +1 (613) 993-3478
Fax: +1 (613) 952-0074
E-mail: George.Forester@nrc-cnrc.gc.ca