Digital halftoning is a technique for converting a continuous-tone image into a quantized image to reproduce it on a digital printing device. Error diffusion (ED) is an algorithm that has proven to be effective for the halftoning process, and it has been widely applied to digital printing tasks. However, in images reproduced using conventional ED algorithms based on the signal processing theory, the texture of objects is often lost. In this study, we propose a texture-aware ED algorithm for multi-level digital halftoning. First, we generate multiple mapped images with different brightness levels through nonlinear transformation. For each mapped image, we adopt a texture-aware binary error diffusion method to obtain multiple halftone images. Finally, we generate a multi-level halftone image from the multiple halftone images. We test the algorithm on an actual printer, compare the results with those of the current raster image processor software and classical ED algorithms, and observe that our algorithm outputs better results.
Digital multi-toning is a technique for converting a continuous-tone image into a multi-tone image for its reproduction with a multi-tone output device. It is becoming important as printers now have the ability to print dots of different intensities. Error Diffusion (ED) is an algorithm that has been shown to be effective for the multi-toning process and has been widely applied to digital printing tasks. However, in the actual printing process, conventional ED techniques for digital multi-toning are often unable to print a sufficiently good-quality image because of physical or mechanical dot gain. In particular, distortion of the contour of printed letters is noticeable. Black letters against a white background appear enlarged, whereas white letters on a black background appear faded. In this paper, we propose an edge-preserving ED algorithm to improve the quality of printed images. We prepare different quantization thresholds between the edge and other regions to allow for stable detection of the edge regions. In addition, we propose a multi-step edge-detection algorithm to avoid printed artefacts. Experimental results using printed images showed that the proposed algorithm improved the quality of printed images in comparison with conventional techniques.