Research On Information Hiding Techniques Integrated With JPEG2000

In cybertimes, it is very important for the information industry to insure the effective transmission and copyright protection of digital information products. As the pivotal techniques correlative to above issues, the coding technique and digital watermarks of multimedia data attract a lot of researchers. Digital watermarking technique is a branch of information hiding. In recent years, research on information hiding techniques including digital watermark has made great progress and its range of application has been expanded gradually.After traversing the information hiding technique and the image coding system of JPEG2000, three information hiding algorithms integrated with JPEG2000 are proposed. The first scheme is a new region-of-interest (ROI) image coding method in which the information hiding technique is utilized. Image coding with regions of interest is one of the important features of JPEG2000. After analyzing the two ROI coding methods of JPEG2000 standard, their shortcoming is pointed out, and then a novel ROI image coding method with utilizing information hiding technique is proposed. In this scheme, some coded information of the background is embedded into the redundant bits of the ROI coefficients. The relative quality of ROIs and background can be adjusted by changing the quantity of embedded information. The second scheme is a robust wavelet-based dual watermarking algorithm for copyright protection of the JPEG2000 coded images. Attack on digital watermarking system destroys different part of the watermark information in various degree. In virtue of the mutliresolution characteristics of wavelet analyzing, a reference watermark is utilized to adjust the reliability of different fractions of primary watermark for reducing the influence caused by the attack. The third scheme, integrated with JPEG2000 baseline system, is an image data recovery technique based on information hiding. Firstly, the wavelet coefficients are modified non-linearly before the quantization stage. Then the relative redundancy of every wavelet coefficients is estimated. The intensity of every embedded point is adjusted image-adaptively according to the estimated redundancy. Finally, the most important part of the code stream is chosen as hiding information that is embedded into high-frequency components of the image. So the unequal error protection scheme for the code stream is realized.