Image Compression Using Wavelet Based Scalable Modeling of Residual Image

• Main Authors: Wan-Shan Lin, 林琬珊
• Other Authors: Wen-Shyong Hsieh
• Format: Others
• Language: en_US
• Published: 2000
• Online Access: http://ndltd.ncl.edu.tw/handle/29533948397683538513

碩士 === 國立中山大學 === 資訊工程學系研究所 === 88 === This thesis is based on the modified JPEG encoding which uses a preprocessing called as Cubic Convolution Spline Interpolation to subsample the original image into lower resolution image, the subsampled image is encoded by JPEG. The modified JPEG can get very high compression ratio, it’s quality is better than the JPEG file which has the same compressing ratio, but it still is not good enough. In this thesis we use the scalable wavelet encoding to encode the residual image, which is the difference between original and compression image. Due to the high compressing ratio and scalablity. We can attach the compressed residual image with modified JPEG compressed image to get the scalable compressed image whose compressing ratio can tun from 125 to 50 and always get better quality than optimal JPEG.