Images have become the most important source for people to acquire information from outside world because of its certainty, visualization, high-efficiency and wide-adaptation. To process images with computers effectively, images needs digitalization, however, because of its time-space characteristics, digital images possess a large amount of data, making it very inconvenient for their storage and transmission, which urgently necessitate compression of images. Researchers have developed many methods for image compression, and fractal image compression, as a new scheme of image compression, has received a great deal of attention and study from researchers all over the world in the field of image compression because of its desirable properties such as fast decoding, resolution independence of decoded image and high compression ratio. However, there's a especially unsatisfying problem in this method: too long encoding time, mainly because of the considerable number of domain blocks to compare with for each range block in encoding phase, which, in fact, prevent fractal image compression from becoming a practical method for image compression, hence fast encoding has become a hot issue in fractal image compression. Existed fast encoding methods are often at the cost of image quality, or can only obtain poor speed-up ratio.In this dissertation, several speed-up techniques for fractal image encoding are proposed, as follows:1. A fast method for fractal image coding based on fractal dimension is proposed. By taking the image to be compressed as a grey-level surface of some fractal brown motion, the fractal dimension of this surface gives out a good description of its texture characteristics. Then the fractal dimension can be used to classifying the blocks in the encoding phase, resulting in a great reduction to number of domain blocks to search for, hence shortening the encoding time dramatically. Theoretically, fractal dimension has powerful ability to classify image blocks, which results in big speed-upratio in fractal encoding; and fractal dimension is a continuous value, which enables us to adjust the quality of decoded images by changing the number of classes. Our experimental results show that, comparing with exhaustive search, when the class number is 25, we obtain 17 times speed-up ratio, with only image degeneration of 1.574951db, at the same compression ratio. Compared with other classification schemes, our method can provide more speed-up and better quality of decoded image.2. A fast method for fractal image coding based on a kind of quadrature. Note that in fractal encoding, each range block is to compare with domain blocks after some grey-level affine map. Hence we first put forward a kind of image quadrature invariant of affine maps above, which can be thought of a generalization of Mario feature. Then this quadrature is used to classifying the blocks in the encoding phase, resulting in a great reduction to number of domain blocks to search for, hence shortening the encoding time dramatically. Theoretically, this quadrature can classify image blocks effectively, which bring out big speed-up ratio in fractal encoding; and this feature is a continuous value, which enables us to adjust the quality of decoded images by changing the number of classes. Our experimental results show that, comparing with exhaustive search, when the class number is 25, we obtain 20 times speed-up ratio, with only image degeneration of 2.08db, at the same compression ratio. Compared with other classification schemes, our method can provide more speed-up and better quality of decoded image.3. A fast method for fractal image coding combining several speed-up techniques is proposed. Based on the principle that range block should share similar intensity distribution with its best-match domain block, we firstly classify all image blocks into shade block and non-shade block, which makes many comparisons between domain blocks and range blocks unnecessary. Then all the non-shade blocks are classif... |