Research On Wavelet Transform-based Scalable Image Coding Techniques

The multimedia applications in a multiresolution environment, such as progressive transmission of image, image database browsing, etc, arise the idea of scalable image coding algorithm, at the same time, the scalable image coding has already become an important target for image coding from now on, because it can save storage space of the image database engrossed by images of different resolutions.Firstly, the scalable coding methods and the results of recent development were investigated, including the basic pattern of scalable coding of time domain, airspace, mass and mixed scalability, the idea and methods of scalable coding technology of frequently-used image compression standard are analyzed, discussed and compared. The wavelet transform on image is introduced; wavelet selection for image is analyzed. The characteristics of the image coefficient distribute, gradation histogram, image itself are emphasized after wavelet transform and the reasonable bases for image compression are obtained after wavelet transform.Secondly, the principle and encoding process of embedded zerotree wavelet coding are investigated, and SPIHT algorithm is analyzed. The bitstream is divided into basic layer and enhanced layer to make it scalable in space by reorganizing, the bitstream of wavelet embedded compression algorithm without increasing extra bits. An optimal bit allocation scheme based on mean square error is given. Simulation experiments demonstrate that the bit of basic layer accounts for a low ratio if compression quality is higher when the image with different ratio of basic layer and quality request is the same, and that the bit rate of basic layer also accounts for a low ratio if the images and ratio are different respectively when the image contains more edge.At last, the application of human visual system on image is investigated to increase efficiency and improve the visual effect of compressed reorganizing image. The tests are done between wavelet basis of V9/3 and Daubechies9/7. After considering human visual system comprehensively and further simplifying SPIHT, a fast image coding method based on human visual system is proposed. The experiment results demonstrate that the algorithm being improved is an effective one on image compression, and is better than other coding algorithm, such as SPIHT, in the recovery quality of the image and coding speed.