Abstract
Permutation coding is a new coding concept proposed by the authors. It has been applied using the discrete cosine transform (DCT) to image compression, achieving high compression ratios at 25:1 and above on grey-tone source images. In this paper, this concept is applied to configure a line (waveform) permutation coder (LPC), using the combination of two methods called the correlation mapping and the permutation innovation as a tool for decorrelating the line-based image data. The LPC is featured in: 1) for being defined in the spatial domain it is implemented about 10 times faster than the conventional DCT-based schemes, while being capable of comparable compression ratios in the range of 6∼10:1; and 2) depending on a sort-based processing manner the LPC, differing from the most of coding schemes which tend to sacrifice image high frequency components in attaining the compression, preserves image edges and contours very well.
