Date of Submission

Spring 2011

Academic Program

Computer Science

Advisor

Robert McGrail

Abstract/Artist's Statement

Given an image or an image Delaunay Triangulation diagram representation, the purpose of this project is to identify a near-duplicate in an image database. This is done by computing the corner points of the image under some default settings, using the Voronoi diagram to adjust these settings in order to get better adjusted corner points for the image, and then using the Delaunay Triangulation of the image to identify whether there are possible duplicates in the database. The project investigates Voronoi Diagrams based on a default Harris Corner Detector as an image segmentation technique for adjusting the threshold for the Harris Corner Detector, as well as Delaunay Triangulation Diagrams as a measure of similarity between images. Ultimately, the project finds image near-duplicates in a database with some likelihood.

Distribution Options

Access restricted to On-Campus only

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

Share

COinS