Fractal Image Coding Based On Irregular Partitioning

Our research base on the current fractal image compression technology with the main purpose of proposing an image coding method which has better subjective quality and coding performance. First, the mathematical foundations for the fractal image compression are discussed. Then, The basic principles, methodologies, general process for fractal image compression and the first atomic fractal image compression scheme proposed by Jacquin are introduced. In chapter four, a new method, IBP, which bases on the irregular partitioning is expatiated and the experiment results are given. Instead of the quad-tree method which partitions the images into square, the method partitions the initial images into irregular blocks according with the content of the image, and therefore the better partitions are obtained. In IBP method, the separate storage and the status code are proposed and integrated into a storage scheme. This scheme is the key to success of IBP method. By exploiting the positional relation among the parameters, the redundancy among parameters is eliminated, and therefore the coding efficiency is improved with no decrease in image quality. This storage scheme provides a new method for parameter storage in image coding. Experimental results prove that by taking full advantage of the characters of images, our method yields a better rate-distortion curve and better subjective quality than the classic quad-tree partitioning scheme.