An Edge-based Image Copy Detection Scheme

Article type: Research Article

Authors: Lin, Chia-Chen | Wang, Shing-Shoung

Affiliations: Department of Computer Science and Information Management, Providence University, Taichung, Taiwan 433. E-mail: {mhlin3,g9371034}@pu.edu.tw

Abstract: This paper presents a novel scheme to detect copied versions of digital images, particularly in an Internet environment based on edge information of images. The edge information is used to generate the signature set for a corresponding image. The edge-based signature sets are then used to compute the distortions between query images and test images, from among which a user wants to find copies. Experimental results confirm that the proposed scheme not only withstands various modifications of an original image but also offers good detection performance in dealing with various rotations, shifting of 17.5% of the width or height of an image, and cutting out up to 45% of the area of an image, which previous research work has found difficult to detect. Therefore, our proposed scheme could be very helpful in detecting copies of digital images transmitted in an Internet environment.

Keywords: Image copy detection, copyright protection, edge-detection, rotation, shifting

Journal: Fundamenta Informaticae, vol. 83, no. 3, pp. 299-318, 2008