Still Image Compression Algorithm

The multimedia(image,audio,vedio) data without compression demands large capacity for storage and transmission bandwidth.Although the technologies of memory,processor and digital communication system progress rapidly,the requirement for storage capacity and data communication is beyond the technologies currently.The data-intensive Web application based on mulimedia arises lately needs a more effictive way of coding signals and images,as well as compression based on storage and communication for those signals.JPEG is the first international standard about still color image and gray image,it is fit for compression of both still image and frames of TV image sequences.In recent years,wavelet transform becomes a hot research aspect of image coding area,for its good characteristic in both time domain and frequency domain,and it has a capability to depict non-stationary image signal and a good performance to fit human visua specialties,if is widely uesed in image coding area and reseach on it has made significant progress.Shapiro put forward EZW(Embedded Zero-tree Wavelet)algorithm in 1993,this method make use of the speciality of wavelet coefficent so that the ouput bitstream has a embedded characteristic,it makes a high compression ratio and good quality of reconstructed image.In this thesis,a comprehensive research for the compression method of still image is addressed, which can be summary as following aspects:1. JPEG baseline based on DCT transform is studied,a method for rating the bitstream is proposed which can achieve gradable stream of different qualities.2. On account of the block-effect of traditional image compression method based on DCT transform and the compression ratio is not ideal,a research of wavelet transform and lifting scheme and coding algorithm by zerotree(e.g,EZW,SPIHT)is addressed.3. Combine integral wavelet transform and algorithm using zerotree,achieve the lossless and lossy compression of image.Propose a improved algorithm of SPIHT and prove it better in both lossless compression ratio and PSNR.