Morphological Pattern Based Approach for Retrieval of Similar Trademarks

In the current scenario of the economic developments in to capture the world market growing role of Trademarks is to attract customers and beat over the imitated/fake products and services. Trademark registration is becoming a critical issue. New trademarks need to be distinct from existing millions of trademarks. This demand searching their similarity out of all trademarks is a typical combination of characters, graphics, images, colors and texture. The number of company and brand is increasing day by day and they all demand a trademark registration for trade practices. In this paper we propose a novel approach to obtain noise free trademark, break it into the character/texture and shape of the trademark logo or design and their relation attributes to build an efficient trademark retrieval tool. We create the input image in a negative image format for ignoring the background and objects reverse intensities values. The paper illustrates the use of morphological opening approach for removing the very small objects which can act as a noise. The work is also implemented to extract the image from the trademark database will be matched further for similar input trademarks on the basis of trademark retrieval process in the database.