A new QIM-based watermarking algorithm robust against multi-image histogram attack

A new attack aimed at DM-QIM watermark was proposed in (Mitekin, 2016). Key recovery results obtained during proposed attack show that for a set of watermarked images it is possible to detect DM-QIM watermark in case if non-local image statistics is utilized. In the present paper, a new modified QIM-based algorithm is proposed for images and video watermarking which shows better robustness against known histogram-based attack. The proposed algorithm is based on modified “correlation-immune” scalar quantizer and does not introduce any correlation between the key bit and the intensity of an individual pixel of the host image. Also, the robustness of proposed algorithm against AWGN is investigated.