Securing Digital Images over the Internet Using a Robust Watermarking Scheme

Abstract

Since 1998, when Chen and Wornell developed a framework known as quantization index modulation (QIM) for characterizing the inherent trade-offs between the robustness of the embedding, numerous QIM schemes have been proposed for securing digital images over the Internet. In general, increasing the quantization steps leads to a good robustness but a poor perceptual quality of watermarked images and vice versa. In this paper, we propose a method for enlarging the quantization steps of a QIM watermarking scheme. This determines the perceptual quality and robustness of the watermarked images. If we choose the quantization steps considering the expected quantization results as well as the original images, we can increase both the robustness and the perceptual quality of the watermarked images. However, larger quantization steps lead to an additional problem, which is false-positive detection. We provide optimal quantization step ranges for the proposed QIM watermarking scheme that minimize the probability of false-positive detection.