Publication: Logo image database retrieval
Loading...
Date
2006-05-01
Authors
Tan, Piet Gek
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Due to the increasing number of registered trademarks, it is significant to have a retrieval system which is able to detect and retrieve the similar or existing trademarks. In this method, images are retrieved by their contents: color, texture, shape, or objects. Thus, the degree of similarity between query images and images in databases can be measured by color distribution, texture distribution, shape similarity, or object presence between the two images. In this project, I proposed three types of retrieval methods: color, shape boundary, and area measurement. A set of 20 color logo images are used as database images and a set of 10 scanned images and example images copied from website as query images. Color-based retrieval is done by computing the average correlation value between query and database images on their respective R, G and B components. Shape boundary-based method is based on Fourier Descriptors representation. The similarity measure is working based on string similarity function. For area measurement method, object areas are calculated as estimation of total on pixels to represent the whole image. Images with least area difference are considered to be more similar to each other. The overall performance of the retrieval system gives acceptable results. A Graphical User Interface (GUI) is also developed in this project for image retrieval purpose.