Technology Research, The Wavelet Coefficients Characteristic Types Of Structure-based Fractal Image Coding (fzbwcs)

Image is the most important carrier among the information intercommunion in people's life, and it is the biggest media containing information. As we all know, the data quantity is very large in a digital image. If the data couldn't be compressed efficiently, it will be a great obstacle to the development of communication and multimedia technologies when high transmission speed and large memory are needed in multimedia communication system. Image compression technology, as a crucial technology in the fields of communication and multimedia, is very important to information technology nowadays and future.In recent years, many new image compression methods have been proposed. One of the most significant advances is the wavelet-based compression scheme that has been put into practice in the areas of compression of static images and motive images as well as a key part in some international standards such as JPEG2000. However, it is still a research hotspot that how to deal with the coefficients transformed from a digital image with wavelet in a more efficient way.The method of fractal image coding is proposed in recent years, and is a new technique used in the image compression. It is a new idea for image compression based on Fractal theory and Iterated Function system. Because of its new idea and high compression rate, fractal coding is abroad focused. It is one of the most futures in image compression at present.When processed by wavelets, the resulting wavelet coefficients of an image represents two characters: the one is that the power of an image is strongly concentrated to low resolution, in other words, there are many zero-trees in the wavelet sub-image; The other is that there is a similarity between wavelet sub-images. According to the above two characteristics of wavelet coefficients, and consider thathuman eyes are sensitive to low resolution distortion but not sensitive to high resolutions distortion, especially not sensitive to that of diagonal directions, we propose an image compression algorithm based on wavelet sub-tree using fractal prediction coding and zero-tree coding.Fractal coding based on the wavelet transform is a coding method aim at the characteristic structure of wavelet coefficients. It makes the best of wavelet coefficients' two characters. First, the image is transformed by the wavelet. And then, we can break all layers' coefficients into sub-tree according as the form of zero-tree structure. Lastly, we set up limit value in order to distinguish the sub-tree, and use fractal coding or zero-tree coding for different sub-trees. In the process of scheme realization, the wavelet coefficients coded by fractal predictive image coding are in high frequency and don't include direct current, so we change the traditional method used by fractal coding. When we use the fractal coding, we change the conventional MSE because there is no direct current in the wavelet coefficients and we use the error emendation matrix to revise wavelet coefficients, and according to the above two characteristics of wavelet coefficients, and consider that human vision system. Theoretical analysis and experimental results presented in the paper show that a high compressing rate is achieved when the quality of image is not declined.To sum up, an image compression algorithm based on wavelet sub-tree using fractal prediction coding and zero-tree coding is a remarkably advanced image compression scheme with good performance. Theoretical analysis and experimental results presented in the paper show that there are a higher compressing rate and quality of the image.