A Low Distortion and Steganalysis-resistant Reversible Data Hiding for 2D Engineering Graphics

A Low Distortion and Steganalysis-resistant Reversible Data Hiding for 2D Engineering Graphics

연구 분야: Strategies

학회: ACM Transactions on Multimedia Computing, Communications and Applications, Volume 19, Issue 2

To reduce the distortion resulting from the large number of crossing quantization cells and resist steganalysis, a reversible data hiding scheme for 2D engineering graphics is put forward based on reversible dual-direction quantization index modulation (RDQIM). The quantization cell index of the host data is first computed, and its distances to the embedding cells in both the left and the right directions are calculated. After that, the data hiding is performed by modifying the data to the nearest embedding cell. To guarantee the reversibility, each quantization cell is further subdivided into three sub-cells, and the source quantization interval of the host data is marked by the index of the located sub-cell. The data extraction is accomplished by calculating the index of the quantization cell where the stego data is in. Meanwhile, the lossless recovery of the stego data is realized by combining the index of the located sub-cell and the relative distance within the sub-cell. Besides, different embedding strategies are adopted for different types of entities to achieve steganalysis-resistant ability. Experimental results and analysis show that the proposed scheme can strike a good balance among imperceptibility, semi-fragility, and steganalysis-resistant ability. Moreover, under the same conditions, the average imperceptibility and the average capacity are, respectively, improved by at least 7.487% and 41.045% compared with the existing methods.