Title: Content-based image retrieval using colour and shape features

Authors: YoungJae Park; KeeHong Park; GyeYoung Kim

Addresses: Department of Computing, Soongsil University, Information Science Bldg., 407, Sangdo 1-dong, Dongjak-gu, Seoul, Korea ' Department of Computer Information Engineering, National Kunsan University, 1170 Daehangno, Gunsan 573-701, Korea ' Department of Computing, Soongsil University, Information Science Bldg., 527, Sangdo 1-dong, Dongjak-gu, Seoul, Korea

Abstract: In this paper, we propose an image retrieval method using colour and shape features. We suggest the content-based image retrieval (CBIR) method using colour and shape information. Using just one type of feature information may be less accurate than using more than two types of feature information. Therefore, many image retrieval systems use many types of feature information, such as colour and shape. We use two features: HSI colour information, especially the hue value, and curvature scale space (CSS) shape information. We search the candidate image from a database that includes the feature information of many images.

Keywords: curvature scale space; CSS; content-based image retrieval; CBIR; histogram intersection; colour; shape features; hue value.

DOI: 10.1504/IJCAT.2013.056023

International Journal of Computer Applications in Technology, 2013 Vol.48 No.2, pp.155 - 161

Published online: 24 Aug 2013 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article