Departmental Bulletin Paper Trademark Image Retrieval using Angular Radial Histogram Approach on Object Region

Moe Zet Pwint  ,  Mie Mie Tin  ,  Yokota, Mitsuhiro  ,  Thi Thi Zin

46pp.221 - 226 , 2017-07-31 , 宮崎大学工学部
Trademarks are valuable things for companies and organizations around the world. Trademarks can represent standard, quality, service and background image of the companies or the organization. Due to the increasing number of business companies and also trademarks, it is important to have a computerized system that can detect and extract the similarity of trademarks because a new trademark must be different from other registered trademarks. Content-Based Trademark Retrieval (CBTR) can deal with matching and detection of similar and infringement trademarks. In this paper, we propose a region-based shape descriptor that combined Angular Radial Approach (ARA) and histogram of object region of trademark image. To confirm the proposed method, the retrieval results are shown according to the rank values and by using the popular dataset.

Number of accesses :  

Other information