Research On Improved Image Compression And Self-Recovery Watermarking Algorithm

The rapid development of Internet technology has accelerated the spread of digital products,but the complex network environment and the easy operation of multimedia itself make the phenomenon of piracy,malicious attacks and illegal tampering more and more serious.The digital watermarking technology that appeared in the 1990s is an effective means to solve this problem and has become a research hotspot in academic and industrial circles.At present,the phenomenon of concealing facts with falsehood and malice is becoming more and more serious in society.Misjudgment of information caused by illegal tampering with original digital products can even lead to serious social accidents.Therefore,some watermarking algorithms based on image authentication and restoration are proposed.In this thesis,some research work has been done on the low quality of the restored image in the existing self-restoring watermarking algorithms.The details are as follows.Aiming at the low quality of decoded image in traditional SPIHT compression algorithm,an improved SPIHT algorithm based on genetic algorithm(Genetic Algorithm,GA)is proposed.The traditional SPIHT algorithm is based on bit-plane coding.It is an image compression algorithm based on the application of classical wavelet transform in image compression.It does not choose the appropriate quantization threshold from different image characteristics.Therefore,the decoding quality of different images is not optimal.In this thesis,through reasoning analysis,we know that distortion is caused by the noise generated by quantizing the wavelet coefficients.Each level plane coding can be regarded as the quantization of the wavelet coefficients by threshold T,and each level T is affected by the initial threshold T0.Therefore,the initial threshold T0 directly affects the quality of decoded image.The improved algorithm optimizes the initial quantization threshold before SPIHT encoding using the function of GA global search for the best individual,so that the optimal initial threshold can be adaptively selected according to the characteristics of the input image under the given output bit rate,so that it can reduce the distortion of decoded image.By testing the images with different texture,brightness and contrast,the algorithm in the thesis performs well in decoding image quality.Compared with the original SPIHT compression algorithm,the experimental data show that the algorithm improves the image decoding quality to a certain extent.Aiming at the problem of low recovered quality under high tampering rate in existing self-recovery watermarking algorithms,a self-recovery watermarking algorithm based on improved joint source-channel coding(JSCC)technology is proposed.Firstly,the image is decomposed into quad-tree by decomposition factor,and the bit-plane is stratified according to the block-classes.The size of each bit-plane is the number of pixels per block-class.Then,the original image is compressed by the improved SPIHT algorithm,and the SPIHT stream is segmented according to the size of bit-plane.In order to prevent the source code information from being destroyed in the transmission process,different RS encoding protection should be applied to each bit-plane to obtain the optimal decomposition result of the corresponding decomposition factor.In addition,the improved algorithm retains the lowest significant bit of the watermarked image when reconstructing the image.This can further improve the quality of image restoration at arbitrary tampering rate.The proposed algorithm in the thesis can allocate more effective bits to high-level bit-planes with high tampering tolerance rate(TTR).At the same time,the bit information used to restore the original image can be embedded into the host image as much as possible.The experimental results show that,compared with the original JSCC watermarking algorithm,the proposed algorithm in this thesis has higher quality of the restored image under high tampering rate,which reflects the effectiveness of high-quality restoration under large-area tampering attacks.