A Comprehensive Study on Content Based Trademark Retrieval System

Abstract

World is emerging with trade practices in the global scenario moving mainly towards the brands both for the products and service equally. Most of the Companies are trying to establish the brand name in the market progressively for the global recognition. As a result, Trademarks play today a very critical and important role. Every Company or Financial organization wants a distinguished trademark for brand name and market viability. Trademark registration and its evaluation for distinctiveness is thus becoming very tedious job for registration offices. Millions of trademarks already registered and millions of applications filed for trademarks registration are aggravating the problem of issuing the trademark certificates. There are different techniques and approaches currently in use for distinctness check for trademarks. The present paper gives the overview of the most popular and appreciated image processing techniques and approaches for the trademark distinctness check. Content Based Image retrieval techniques are widely used for that purpose and some other approaches like shape and texture based similarity finding techniques are also used. In the present paper, our approach is to summarize the most widely used techniques for the trademark distinctness check.

Fuzzy Clustering of Image Trademark Database and Preprocessing Using Adaptive Filter and Karhunen–loève Transform

“…The color feature has been extracted using color histogram for dominant red, green and blue color channels (RGB). 13,15 The database images are partitioned using fuzzy c means clustering into three clusters. The initial testing is done using 500 database images.…”

Content based image retrieval system based on watershed transform for trademark images

“…This algorithm is to enable the implementation of watershed transform using a simple approach based on breadth algorithm. (12) where k=2 represents the second biggest object from the segmentation result, with n number of pixels composing the region and m number of boundaries surrounding the object. While p b represents pixels in the boundaries and p c is the center of the region object.…”

“…In this field, many different techniques have been proposed and implemented such as knowledge-based system [5], geometry representation and spatial similarity [6], contour skeleton segmentation [7], morphological pattern [8], size feature [9], relevance feedback [10], 3D projection [11] and the utilization of shape, color and texture [3,12,13]. Commonly the main problem suffered by the most existing CBIR system above is in the image analysis part which deals with the function to define the content of the image.…”

“…Both features are computed from the second biggest region produced by segmentation process, assuming that this region is the dominant object while the biggest region is the background. To proceed with those features, the following calculations are held as Equation12.…”

Content based Trademark Retrieval by Integrating Shape with Colour and Texture Information

“…Before proceeding with this paper work an analysis of the survey done on various CBIR systems by Ranjeet Kumar et al [10] has been made to get a better insight into the work already done in this field.…”