Image tampering detection based on feature consistency attention
by Junlin Gu; Yihan Xu; Juan Sun; Weiwei Liu
International Journal of Information and Computer Security (IJICS), Vol. 23, No. 1, 2024

Abstract: Recently, the development of computer image technology makes image tampering more and more convenient, causing a large number of image tampering accidents. The existing scheme uses manual features and depth features to analyse the forgery traces, and has achieved good results. However, the existing schemes lack the analysis of essential traces and have defects in generalisation performance. In this paper, an image tampering detection scheme based on feature consistency attention is proposed. The inconsistency between the real region and the background region is used to improve the detection ability of the algorithm for unknown images. The scheme uses the feature extraction module to extract the deep semantic features of the image, and then calculates the feature correlation between the tampered region and the background region to maximise the correlation within the region and minimise the correlation between the background region and the tampered region. The scheme can learn the common traces of tampering process, which is expected to achieve better generalisation effect. Experimental results show that the proposed scheme is superior to several existing schemes in detecting tampered images.

Online publication date: Mon, 19-Feb-2024

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Information and Computer Security (IJICS):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com