Hybrid Image Compression Encoding Based On IFS Theory

Fractal image compression (FIC) is developed in the last 10 years and takes advantage of self-similarities which exist in the images commonly. There is a certain kind of self-similarity in natural images. Fractal image compression can be achieved so long as a group of affine transformation whose fixed point is the approximation of the original image is found. FIC realize high compression ratio because the code consists of parameters of the iterated function systems instead of the pixels of the image. FIC became a hot research in the field of image compression encoding.Based on fractal theory, the following works are done in order to improve the compression ratio and reduce the huge encoding time of fractal image compression:Firstly, this paper realizes a hybrid image compression method based on spatial correlation and evolutionary algorithm. The searching space of the current range block is constrained to the eight neighborhoods and its extended blocks, which greatly speed up the encoder; hybrid genetic algorithm is operated to explore more adequate similarities if the local optima are not satisfied. Experimental results show that encoding time and the compression ratio have both improved while the quality of retrieved image declined slightly.Secondly, an improved fractal image compression method is designed. It improves the known fast encoding algorithm based on local variances. According to the statistics of image blocks, affine transformation is used for encoding if the variance of a range block is less than the predefined threshold, while hybrid neural network is used if the variance of a range block is bigger than the predefined threshold to reduce the matching error. The method not only reduces the matching error for subblocks, in which the pixel values fluctuate greatly, but also overcomes the disadvantages such as slow convergence and local convergence of BP network used in image compression. Experimental results show the effectiveness of the method.Thirdly, based on no search and neighborhood fractal image compression, an improved fractal image compression method based on progressive/segmentation neighbor searching strategy is realized. The searching space is constrained to five neighborhoods of the current range block, and the optimal searching extends progressively. If the solution has no improvement, the subblock will be segmented into four parts. Redundancies in traditional quad-tree partition storage are compressed. Experimental results show the effectiveness of the method.