A Binary Watermarking Algorithm for Color Images Based on Structure‐Preserving Quaternion QR Decomposition

ABSTRACT We introduce a robust blind binary watermarking algorithm. The method starts by dividing the host color image into nonoverlapping small blocks which are quaternion matrices. For these matrices, a complex structure‐preserving QR decomposition algorithm is proposed to implement the embedding and extraction, which has enhanced speed and accuracy compared to other QR decomposition algorithms. To further improve the watermark's invisibility and robustness, we select the positions in the decomposed matrices with the highest normalized correlation coefficient (NC value) and the color channel with the highest peak signal‐to‐noise ratio (PSNR) for embedding the watermark. The experimental outcomes demonstrate that our proposed watermarking algorithm achieves the highest computational efficiency in embedding, attaining an optimal Bit Error Rate(BER) of 0 and a NC value of 1 in attack‐free scenarios, thereby outperforming existing methods. Under the attacks such as Gaussian‐blurring, sharpening, salt & pepper (S&P) noise, JPEG‐DCT, and rotation, our scheme achieves a higher NC value compared to other existing methods. Overall, the results show that the proposed watermarking algorithm offers high time efficiency and surpasses other algorithms in both invisibility and resistance to the specified attacks.