Fast Fractal Image Compression Method Of Research

Recently , there are hundreds of image compression methods, but the effect of compression, compression ratio and the time of coding and encoding can not satisfied with the needs of information era. So image compression needs more development. Because the traditional coding methods have generally been formulary, numbers of new methods come forth in recent years with the development of many new theories. Fractal image coding is one of the new coding methods. Fractal image coding possesses the novel ideas, the great development potential and when it's compression ratio reaches to 10000:1 on the conditions of people's intervention, the visual effect of encoded image is very good, so fractal image coding is a method of wide development prospect. Fractal image coding is based on the collage theory. According to the given image, it can search out a group of contractive mappings in order to make the attractors of Iterated Function System approach the original image, and then the parameters are memorized. The encoding procession is based on the Iterated Function System, which is composed by the memorized parameters. The encoded image comes out through being iterated for several times.Fractal image coding is a coding method that sprang up newly in the 1990s last century, it has enormous development potentiality Shown in the following several aspects concretely:1) The compression procession, which thinks over the relationships between some and during the part, as well as the relationships between some and the whole, is based on an image of self-similarity. There are abundant of self-similarity geometry in the world, so fractal image coding is in a very large range to suit.2) The resolution ratio of encoded image has nothing to do with the size of original image. The size of the encoded image can be shrunk or enlarged as much as possible, furthermore this can not destroy the elaborate structure of the encoded image.3) Fractal image coding possesses very high PSNR as well as the good visual effect under the conditions of high compression ratio, which other methods can't compare with.Fractal image coding and encoding has been developed for over 10 years, however its potentiality has not been excavated out, for example: its compression ratio isn't so high, the coding time is not so short and it is not the very idea to the compression effect, so recently in the image coding methods, fractal image coding can not occupy the leading position.There are a lot of places in the fractal image coding less than satisfaction, the slow coding speed is one of the factors that prevent it from being practicability, this paper analyses the factors that affect the speed of fractal image coding , and proposes a novel idea by using the reformed variance (tentatively) to improve image fractal compress performance.The most time-consuming part during the coding is to search out the matching block. In order to improve the coding speed, we must to reduce the searching space to make the best matching block exist in a certain small range. Classification is one of the methods for people to improve the coding speed. The effect of classification will alter as the different of criterions. The criterion based on variance or on entropy has much contribution to advance the speed of fractal image coding, yet there are many shortcomings about the two criterions. The advanced variance of image block will not change as the existence of the IFS, and it can gain the matching block with high precision compared with the other criterions. So it is possible to improve the speed with theadvanced variance criterion.The hard core of the fractal image coding based on advanced variance is that: to classify the Range block and Domain block before matching according to the static feature, and to search the Domain blocks which are in the same sort. This can improve the coding speed highly on the conditions of not to depress the quality of encoded image. In the course of realization, there are a lot of advanced measures in the coding method, which including predigesting the error function, design the window and so on.The experiments show that the searching time of the fractal image coding based on advanced variance is far smaller than that of the coding based on the variance or on entropy, furthermore, the PSNR of the encoded image is better than other fractal based on the static feature.In sum, the fractal image coding based on advanced variance is a method of good performance. It can improve the speed of coding, as well as the quality of the encoded image.