Robust Double Domain Watermarking for Binary Document Image
Xin Zhu · Chinese Journal of Computers · 2014
This paper presents a double domain watermarking(DDW)method for binary document image.In this method,the watermark patterns are generated as signals in the discrete cosine transform(DCT)domain,and modulated with the embedded message.The modulated signal is then transformed into the spatial domain and adapted to the characteristics of human visual system by adjusting the localized embedding strengths before embedding.Watermark extractions can be performed by applying a correlation based detector in either of the two domains.The effects of perceptually shaping and the number of the flipped pixels on watermark decoding are also analyzed carefully.The DDW method is combined with an image normalization procedure,which is developed particularly for document image through skew detection and character segmentation,in order to resist synchronization distortions.The experimental results show that the presented scheme manifests the good robustness against common image manipulations,geometrical attacks as well as print-scan process.