Share Email Print

Proceedings Paper

Security threat to media security applications based on scale-space feature extraction
Author(s): Chao-Yong Hsu; Chun-Shien Lu; Soo-Chang Pei
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Scale-space image feature extraction (SSIFE) has been widely adopted in multimedia security and other applications due to its powerful resilience to attacks. However, the security threat to SSIFE-based media security applications, which will be addressed in this paper, is relatively unexplored. The security threat, composed of a constrained-optimization keypoint inhibition attack (KIHA) and a keypoint insertion attack (KISA), is specifically designed in the proposed method for scalespace feature extraction methods such as SIFT and SURF. The principle of KIHA is to make a fool of feature extraction protocols in that the detection rules are purposely violated so that no local maximum can be found around in a local region. On the other hand, KISA is designed to create the false positive problem. Our method is evaluated and compared with Do et al.'s method (ACM MM'10), which also figures out the weakness of our previous work (ACM MM'09). In addition, our proposed security threat is applied to an image copy detection method operated on a web-scale image database for performance evaluation.

Paper Details

Date Published:
Proc. SPIE 8303, Media Watermarking, Security, and Forensics 2012, 830303; doi: 10.1117/12.908434
Show Author Affiliations
Chao-Yong Hsu, National Taiwan Univ. (Taiwan)
Institute of Information Science, Academia Sinica (Taiwan)
Chun-Shien Lu, National Taiwan Univ. (Taiwan)
Soo-Chang Pei, Institute of Information Science, Academia Sinica (Taiwan)

Published in SPIE Proceedings Vol. 8303:
Media Watermarking, Security, and Forensics 2012
Nasir D. Memon; Adnan M. Alattar; Edward J. Delp, Editor(s)

© SPIE. Terms of Use
Back to Top