Secure and Robust Watermarking Using NSST-LWT-DCT-SVD for Copyright Protection

Abstract

The ongoing issue of unauthorized image distribution poses a serious modern challenge. This research introduces an innovative dual image watermarking technique that determines a potential embedding area in a small 4×4 DCT block and employs a hybrid approach of NSST-LWT-DCT-SVD to ensure efficient copyright protection. The watermark images are integrated using the red and green channels. These channels are then transformed using NSST-LWT to extract the low-frequency sub-band. Following this, the low-frequency section is divided into 4×4 blocks. The Human Visual System (HVS) is then used to select the appropriate blocks, which are transformed using DCT. Certain coefficients from the middle frequency of the DCT are chosen for decomposition with SVD. The singular values are then employed for embedding the watermark bits. Experimental results indicate that the proposed technique is a viable option for image watermarking, achieving average PSNR and SSIM values of up to 45.9 dB and 0.9962, respectively, while closely approaching a value of 1 for NC performance.