LWT Based Robust Color Image Watermarking Using QR Factorization and DCT

This paper proposes a novel robust blind color image watermarking technique that combines entropy analysis, lifting wavelet transform (LWT), QR factorization, and discrete cosine transform (DCT) for copyright protection. The RGB components of the color image are first separated, and the R component is then subjected to the 2-level LWT. Following DCT is applied to the 2-level LWT's high-frequency sub-band, the DCT coefficients are separated into 4x4 non-overlapping blocks. Each chosen block is then subjected to QR factorization, and the watermark is embedded in the first row and first column element of the upper triangular matrix. Numerous simulated tests show that the presented scheme is highly imperceptible and robust to attacks. The presented watermarking scheme performs better than the others in terms of invisibility, according to comparisons with comparable schemes.