Histogram Preserving QIM-Based Watermarking
Chang‐Tsun Li, Yue Li · Warwick Research Archive Portal (University of Warwick) · 2012
Quantisation-based watermarking schemes are known to be strongly robust to noise and widely applied for image authentication. However, such schemes tend to create specific patterns in the histogram of the watermarked image, thus opening security gap for histogram analysis and attack. As such, a secure quantisation-based watermarking scheme should be histogram preserving. That is to say the original statistical characteristics of the histogram should be preserved in the watermarked image. In this paper, we propose a histogram preserving quantisation index modulation (HPQIM)-based watermarking, which can preserve the original statistical characteristics of the histogram of the original image after watermark embedding. This scheme originates from quantisation index modulation (QIM)-based watermarking scheme, but highly increases the security of QIM-based schemes. Meanwhile, the proposed HPQIM-based watermarking scheme can also achieve a lower embedding distortion than the QIM-based watermarking schemes.