本文提出以小波轉換為基礎之一套新的影像編碼技術，是利用經由小波轉換係數 值有接近於灰階度零的效果，及其擁有正負二值的類型將小波係數以四個係數為 一組的小波係數組做為編碼的依據。這種設計方式可避免因訓練產生區域性的最 佳化。其訓練方式是利用歐幾里得最小距離，經由反覆訓練產生最佳的小波係數 組。而這些小波係數組再利用Huffman編碼法進行編碼。實驗結果證明本方法與 傳統的小波轉換編碼技術比較，可達到更高品質、更低位元率的影像壓縮。 Image compression coding has high potential on some applications in our life. It can be used in a wide range of application. Such as the storage of the digital camera, multimedia transmission in network, image database, etc. As a result, in the recent years many researchers have proposed relevant methods and systems for image compression. In this paper, we proposed the image compression system is based on Discrete Wave-let Transform (DWT). The wavelet coefficients is divide into Wavelet Coefficient Set (WCS) that will be coded. This system is composed of three parts: In the first part is image spatial transform. It extract wavelet coefficient by mutual orthogonal feature with wavelet function and scaling function of DWT. In the second part is training WCS. It can be replace original coefficient by trained coefficient set. In the third stage is coding method. It is coded by huffman coding to replace with original one to reduce to the bit rate of compression. In this system, we use 3 popular testing graphics for compression. Compare with the other compression method in PSNR's and bpp's, it has the better effect, too. We use subband with rim detection by DWT in application. It put to use the advantage sufficiently, and extract better effect, too.
銘傳大學展望新世紀國際學術研討會資訊組論文集=Proceedings of INTERNATIONAL ACADEMIC CONFERENCE FOR THE NEW MILLIEUM，頁521-527