Suprathreshold stochastic resonance and maximizing network for watermark detection

摘要

We propose a novel method that improves watermark detection performance, based on suprathreshold stochastic reso nance (SSR) with a maximizing network. The detection pertormance is measured in terms of correlation. The proposed method has been tested on different gray-scale images, and we found that an original image is not required for watermark detectbn, so this method is blind. We improved the correlation between an original watermark and the SSR with maximizing network-based discrete wavelet trans form coefficients of the watermarked image. Our experimental results have been compared with the different existing techniques and were found superior in terms of correlation and ratio of correlation to thresh old.