摘要 |
<p>The multidimensional multi-valued color image compression and decompression method enables an image to be compressed and decompressed at a high efficiency, particularly in a system in which the pixel or frame correspondence is not guaranteed on both the compression and decompression sides. In the method, a specific luminance boundary is obtained by tracing a pixel contour for each specific luminance threshold previously determined for an image luminance function; the specific equi-luminance line passing-point coordinates are shifted to a central position of each of the pixels for constituting the obtained specific luminance boundary; the feature points satisfying a specific condition are obtained on the specific equi-luminance line; and the positions and the luminance values at the image feature points are transmitted or recorded as information obtained by compressing the original image information at a high efficiency, in order to restore an image by decompression. Further, on the decompression side of the image reproduction, the bright region whose boundary is determined by the equi-luminance line is marked by a bright symbol, and the dark region whose boundary is determined by the equi-luminance line is marked by a dark symbol, both for each luminance threshold for obtaining a specific equi-luminance line, so that a binary mask can be formed by the bright and dark symbols; the equi-luminance line is approximated into a polygonal shape; and the luminance values at the pixels other than the feature points are decided on the basis of the interpolation plane decided by a plurality of feature points locating in the vicinity of the equi-luminance line polygonal-approximated by the lines connecting the feature points. <IMAGE></p> |