Video Compression Based On Wavelet Transform And Generalized Finite Automata

With the rapid development of multimedia technology in recent years, the digital video transmission plays a very important role in communication field. Digital video generally include a great deal of information, so the compression technology becomes the key of the research. On this occasion, it inevitably becomes one of the most important directions in the field of communication and signal processing to seek after the effective video coding algorithm. The technique of Finite Automata has a potential capability of improving coding performance. However, the development of the video coding based on FA has not been perfect. So in the paper, an in-depth study of video Generalized Finite Automata (GFA) coding algorithm is conducted. The major jobs are described as follows:(1)Basic knowledge of video and main methods of video compression is introduced. At first, the development of video compression coding is reviewed simply; then the estimation methods of video quality are discussed, providing an academic preparation for image compression coding.(2)The wavelet-base coding is recommended in image compression and video compression. The experiment is performed with SPIHT (Set Partitioning in Hierarchical Trees) algorithm.(3)The background and the development of the fractal and the finite automata are reviewed. Besides, the basic conceptions and symbols of the fractal and the finite automata are introduced in details. The most important aspects of this section are the detecting of video images'self-similarity and the finite automata for multiresolution image.(4)According to the principle of self-similarity, the GFA compression algorithm is designed. In the proposed scheme, a video sequence group of pictures (GOP) is represented as an overall binary image by bitplane modeling the coefficients of the video sequence within subbands using a generalized Gaussian quantizer, adaptive to the coefficient statistics of the video sequence. The inter-frame (motion), interband (frequency) and inter-bitplane (spectrum) similarities are then optimally explored in the bitplane model, to form a compact GFA representation of the video sequence. Finally, all the transitions (edges) and the leaf nodes in the GFA representation are entropy encoded into a scalable bitstream. It is an effective scalable video coding scheme. Numerous multimedia communication applications, previously unpractical, would be envisaged with the low bitrates.Finally, the research work involved in the thesis is summarized and the future developments in GFA method are forecast.