Multiple spatial channel printing

Abstract

Abstract A multiple spatial channel technique is described that can produce near continuous-tone output quality from an output device having a large, dilute dot channel, referred to as the L (lowpass) channel, and a small, concentrated dot channel, referred to as the N (noisy) channel. The new process described here overcomes contouring artifacts and produces vastly improved images over techniques that use only a single dot type. Given a matrix of image luminance values (raster), two images are computed for printing with the two channels of the printer. The two images are computed simultaneously with a two-pass algorithm, one pass horizontally and one pass vertically. Initially, the L channel image is set to be equal to the input image. Each adjacent pair of pixels in the original image is compared by calculating an edge contrast strength. If the ratio of the first pixel level to the second pixel level is larger than a threshold, the L channel pixel that corresponds to the second image pixel is set to the first pixel level. On the other hand, if the ratio of the second pixel level to the first pixel level is larger than the threshold, the L channel pixel that corresponds to the first image pixel is set to the second pixel level. This procedure erodes the dark sides of edges that have contrast above the specified threshold. Similarly, during the computation of the N channel image, each pixel in the input image is compared to its corresponding L channel image pixel. The corresponding N channel image pixel is set to be the ratio of the input image pixel to the L channel pixel. The technique is extensible to multiple sizes of large dilute dots and can be combined with image resolution techniques. The separation technique can also be used for display, and has several advantages over the prior art including low noise, low computational expense and low bandwidth requirements.

Inventors

• 👤 D. Amnon Silverstein

Bibliographic Details