The data compression system and - a method by means of sub-sampling and preferably an adaptive accounting for the transmission of images to a printer

摘要

Pixel-based image data (11) are subsampled (13) for transmission (14') to a printer (Fig. 2) and later bilinear interpolation (31) in the printer. The transmission system (Fig. 1) evaluates (15, 23) the error (24) that will remain after interpolation. In response to this evaluation, the transmission system also forms (25) trim data (26) for later application to the interpolated subsample (32, Fig. 3c) to remove a significant part of the error (24). Preferably trim data points (be'', bf'', df'') are selected based on relative importance (Fig. 3d) of the associated correction - specifically, size of the error that will remain after interpolation - but also the amount of trim data is controlled to maintain a significant advantage in overall transmission time, relative to transmitting the entire image-data set. The amount of trim data is held roughly to a certain fraction, about five percent, of the number of points in each data block (e. g., swath), essentially independent of the error evaluation. Preferably this controlled amount of trim data is padded out with zeroes (ab'', ad'' etc.) to form a complete trim-data array - which is then run-length encoded (Fig. 3f). The encoded trim data (28) and the subsample (14', Fig. 3b) are sent to the printer. (Preferably pure-text portions Ä19Ü are sent separately.) The printer applies (35) the trim data (28) to the interpolated subsample (32, Fig. 3c) to obtain a better approximation (36, Fig. 3g) to the original image data (11, Fig. 3a). IMAGE