Image Compression based on theories of Gaussian Frame and Feature Extraction

• Main Authors: Chen, Jia-Shuang, 陳家祥
• Other Authors: Shin Shaw-Jyh, Hsin Chengho
• Format: Others
• Language: zh-TW
• Published: 1998
• Online Access: http://ndltd.ncl.edu.tw/handle/99092307705596826217

碩士 === 逢甲大學 === 電機工程研究所 === 86 === The purpose of this thesis is to develop a static image compression scheme that is based on theories of Gaussian frames and feature extraction. The primary motivation of this research is to solve the problem of destroying salient image features in traditional image compression methods for low-bit-rate coding. The developed image compression scheme is composed of an encoder and a decoder. In the encoder, image features are extracted first, and then both theparameters of the extracted image features and the low frequency components ofan image are coded. In the decoder, the high frequency components of an image are predicted by the decoded parameters of the image features. Hence, the original image can be reconstructed by adding the predicted high frequency components and the decoded low frequency components. The initial simulation results show that the proposed compression scheme retains most of the salient features for low-bit-rate coding.