Bi-level rendition of images containing text, screened halftone and continuous tone

The authors propose a novel adaptive bi-level quantizing technique for images containing text, screened-halftone area, and continuous-tone area on the same page. Screened-halftone area is first separated from nonscreened-halftone area, and then text area separation is done for nonscreened-halftone area. In the separation step, the wrong separation probability decreased greatly from that of the conventional algorithm by utilizing hysteresis incorporated in the criteria dealing with image segmentation variables. In the proposed bi-level quantizing step, error diffusion is used; however, an additional technique is employed in which three parameters, i.e. error feedback, ordered dither merging, and averaging in screened-halftone area, are also controlled according to the image segmentation variables