When designing steganographic schemes, matrix embedding is an efficient method for increasing the embedding efficiency that is de- fined as an average number of bits embedded via per change on the cover. Random linear code-based matrix embedding can achieve high embedding efficiency but cost much in computation. In this paper, we propose a method to increase the embedding speed of matrix embedding by extending the matrix via some referential columns. Compared with the original matrix embedding, the proposed method can exponentially reduce the computational complexity for equal increment of embedding efficiency. Experimental results also show that this novel method achieves higher embedding efficiency and faster embedding speed than previous fast matrix embedding methods, and thus is more suitable for real-time steganogaphic systems.

Wet paper codes (WPCs) are designed for steganography, in which the sender and recipient do not need to share the changeable positions. In this paper, we propose theN-page construction for wet paper coding, which can generate a family of WPCs following the upper bound on embedding efficiency from one single WPC. The Paper Folding method, one of our previous methods, is a special case of theN-page construction withN=2k. We deduce recursions for calculating embedding efficiency ofN-page construction, and obtain explicit expression on embedding efficiency of 2k-page construction. Furthermore, we derive the limit of distance between the embedding efficiency of 2k-page construction and the upper bound of embedding efficiency asktends to infinity. Based on the limit, we analyze how the embedding efficiency is influenced by the proportion of wet pixels (wet ratio) in the cover, showing that embedding efficiency only drops about 0.32 as the wet ratio increases to 0.9999.

In this paper, a new embedding strategy for DWT-based watermarking is proposed. Different from the existing watermarking schemes in which low frequency coefficients are explicitly excluded from watermark embedding, we claim that watermarks should be embedded in the low frequency subband firstly, and the remains should be embedded in high frequency subbands according to the significance of subbands. We also claim that different embedding formula should be applied on the low frequency subband and high frequency subbands respectively. Applying this strategy, an adaptive algorithm incorporating the feature of visual masking of human vision system into watermarking is proposed. In the algorithm, a novel method to classify wavelet blocks is presented. The experimental results demonstrate that the watermarks generated with the proposed algorithm are invisible and robust against noise and commonly used image processing techniques.