| Summary: | 碩士 === 國立高雄大學 === 電機工程學系碩士班 === 104 === Images are important media because that can represent the information precisely. Due to the massive size of image files, how to perform compression effectively is the major task for tablets or smartphones. There have been lots of international standards for fulfilling the compression procedures. In this thesis, unlike using conventional techniques, we employ the recently developed scheme named compressed sensing for data compression. In addition to the effective compression, we also aim at the robust transmission of compressed sensing signals over lossy channels. We employ two techniques for protecting compressed sensing signals, namely, multiple description coding and error control codes, for fulfilling the requirement of robust transmission. Simulation results have presented the alleviation of reconstructed image qualities for the use of correlations with the coefficients in multiple description coding, and the inherent characteristics of error correction with error control codes. Proposed methods also have the potential for the integration into practical implementations in the future.
|
|---|
