A Confidential Communication-Oriented Information Hiding Algorithm Based on GHM Multi-Wavelet and DCT

Structure and energy properties of carrier are considered as the most important factors to information hiding performance. This paper develops a new method to analyze energy and structure characteristics of digital images in order to preprocess the carriers. We adopt GHM multi-wavelet transformation to decompose the cover image into sub-images with different energy level. Then we obtain the embedding regions which are expressed as numerical range by discrete cosine transform. We embed data with different robustness in different regions. Then we get the stego image which is rarely changed in energy and structure properties compared with the original cover image. Experimental results indicate that the invisibility and robustness can be increased separately by 26.87% and 19.25% averagely, and the ability against steganalysis such as RS and Higher Order Statistics based on wavelet coefficients can be improved. Moreover, our algorithm has excellent sensitivity of image processing.