Image Coding Technology Research Based On Multiscale Geometric Analysis

Image compression coding , As modern communications, media storage, data distribution, multi-media key link, has been most important in data processing areas. In order to effectively store, process and transmit large amounts of data, and enhancing the computer's storage capacity, processing speed, and efficiency of information transmission image, digital image compression coding technology in the field of multimedia information processing has shown that more and more important position. This study was carried out in this background.In image coding area, Multi-scale and time-frequency analysis theory is the core technology and hot spots. Wavelet from be proposed to now is the classic method in this area. However, due to the inherent limitations in its theoretical, making it in dealing with high dimensional data space it is powerless, against such defects in recent year based on the wavelet transform it has developed many new analysis tools In my paper,we focuses on two new multi-scale and time-frequency analysis tools: contourlet transform and bandelet transform. analysis of both Theory and transform their own identity, its image in the digital encoding, digital watermarking, such as the effective application of research carried out more further studies.We known that wavelet transform for image is difficult to capture the characteristics of high-dimensional accurately, in this paper we use bandelet transform and Contourlet transform to process image combined with experiment,so we proposed three algorithms as following: 1, A new image coding using coutourlet transform and human visual system is proposed in this paper. Firstly, the wavelet transform is applied to original image, and the directional filter banks(DFB) are employed to middle and high frequency subbands; Then, the high frequency coefficients are weighed according to the human visual system; Thirdly, the new extended spatial orientation tree are constructed; Finally, the encoding scheme is realized by the idea of SPIHT. 2, in this algorithm we has improved on the second generation bandelets transform, in order to overcome some disadvantage effectively about bandelets transform in image coding, such as excessive simpleness of quantization and high time complexity and so on. .Fist of all, this paper carries bandelets transform though the origin images, and then carves up direction block by using the relativity between wavelet scale quotiety. Next, the article computes the best threshold value T under fixed code rate. Finally, aiming at the quotiety distributing characteristic after bandelets transform, the article makes the maximum value quotiety as the certer, and scans extending around in rectangle scan form.The experimental results show that the above proposed algorithms in the three image coding and digital watermarking field are better than classic, universally applicable image processing algorithms, and has a strong versatility and adaptability, it has good application prospects.