Research On Image Coding Based Discrete Cosine Transfom And Wavelet Transfom

The transfer and storage of image is an important area in modern information communication. Under equal communication capacity, if image data can be compressed before being transferred, the amount of transferred data can be made smaller, so as to enhance communication capability. At the same time, the same storage space can store more information. Therefore, people are taking the efficiency of image compressing for more consideration.The image coding algorithms based on DCT and wavelet transform are emphatically analyzed in the thesis, which are the two most popular image coding methods in recent years. DCT coding has been widely used in coding standards such as JPEG. However, the image codings have been advanced to a new stage since wavelet's embedded zerotree and SPIHT are successfully applied, which give a stern test to conventional DCT coding.Whether DCT will be displaced wavelet transform? These questions have been puzzling people. So a comparison between image coding methods based on DCT and wavelet transform is presented in the thesis: After reorganizing or layering DCT's coefficients, DCT's improved algorithm are obtained. Moreover, DCT's property similar to wavelet's is presented. Then JPEG based on DCT is compared with JPEG2000. After that, we will draw some conclusion after some experiments by comparing image coding algorithms based on DCT and Wavelet. PSNR will be the parameters for comparing. Based on experiments , DCT still hold broad application prospects in the image compression. By the comparison, it is fact that primary factor for the transform domain coding lies in the quantizer and the entropy encoder, instead of distinguishes between DCT and DWT. DCT and wavelet transform have their own advantages respectively, in the application we can carry on reasonable choice according to different need. Finally, it is proposed that DCT and wavelet transform be combined together for image-compressed. This algorithm can improve compression ratio further and get fine compromise between CR and PSNR.