Image Authentication System Using Image Hashing with Color Filter Array

Abstract

AbstractThe availability of sophisticated multimedia editing tools makes image authentication a challenge in the modern age of multimedia communication. We suggested a hashing approach for image authentication in this paper. The proposed system extracts invariant vector distance features form a secondary image which is constructed through ring partitioning of the quaternion image. In parallel, the color filter array (CFA) interpolation coefficients are calculated from the CFA image, which has been used for identification of tampered regions. The concatenation of invariant vector distance and CFA coefficients construct the final hash. The results of experiments indicate that the approach proposed is robust in maintaining content against brightness/contrast adjustment, addition of noise, and arbitrary rotation etc., and identified the tampered regions. Compared to the existing image authentication methods, our approach provides better performance for forgery detection.KeywordsImage authenticationCFA interpolation coefficientsRing partitionInvariant vector distanceCentral orientation informationRobust image hashing