Logo extraction using dynamic stochastic resonance

This paper presents a novel technique for watermark (logo) extraction from the distorted watermarked images. The proposed technique is based on the combined discrete wavelet transform (DWT), discrete cosine transform (DCT), and dynamic stochastic resonance (DSR). In this technique, DCT transform is applied to the DWT coefficients of the watermarked image. Later on, variance is calculated from the DCT coefficients, and it is used to tune the DCT coefficient with the help of dynamic double-well system parameters that causes resonance to DCT coefficients. The experimental results have been compared with the existing techniques and were found superior in terms of peak-value-of-signal-to-noise-ratio (PSNR) and correlation coefficient.