Dct-based Image Coding

Image compression encoding is one of the key techniques in modern multimedia and communication field. Among kinds of compressing methods, discrete cosine transform (DCT) is the core of most international image compressing standards.But, there are some obvious disadvantages in traditional DCT transform, such as long encoding and decoding time costing, blocking effects and so on. In this dissertation, some researches are developed to overcome these disadvantages and to improve the image compressing quality with fractal Peano-hilbert scanning curve.Firstly, the reducation of DCT operation time could be acquired according to a new algorithm about Block-based DCT transform. After an input image is partitioned into some same size blocks, each block will be sanned and rearranged to an one-dimension vector without direct 2D-DCT. Then, 1D-DCT can be implemented and operation time will decrease.Secondly, an attention is payed on the improvement of hierarchical discrete cosine transform in this dissertation. The "blocking effects" can be reduced and the coding rate can be increased with hierarchical DCT, zerotree quantizer and encoding according to the Peano-hilbert scanning sequence.Experimental results show that image compress quality is well and image compress/encoding time decreases notablely.