Studies On Digital Halftoning Based On Noise Properties

The theme of this paper is form "the Tenth Five Plan" military electronic project "Consolidated Inkjet Graph Plotter" and the project "Research of RIP Software Key technologies of Color Output Device", which is appraised by the natural science fund of ShanXi province. RIP (Raster Image Processor) software is very important for the performance of inkjet printer. Digital halftoning is one of the key technologies of the RIP software, directly determine output images quality, so its position in the printing system is very important.On the basis of integrated analysis of the conventional digital halftone technique characteristics, we use the printer's model and visual model to find the reasons of causing the printer's distortion and the methods to solve these problems.Given the low-pass nature of the human visual system (HVS), blue-noise halftone technology creates patterns visually appealing because the spectral components of the pattern lie in the regions least visible to the human viewer, and the artificial textures are not apparent to the viewer. The green noise halftoning have these blue noise advantages, besides, it can be used in the harder condition, so the green noise have more halftone robustness than the blue noise. This paper mainly studies blue-and green-noise digital halftoning methods in spatial and spectral domain, the blue-noise can be seen as a special case of green-noise, once there is only one minority pixel in the cluster, it becomes a blue-noise pattern. Based on these theories, we can deduce the error diffusion with output-dependent feedback method that can control the dot-gain distortion well.At last, we present the generic way to create a blue-noise mask and study a multi-gray blue-noise digital halftoning arithmetic, based on that, we present a new green-noise halftoning arithmetic with adjustable cluster size, this method can effectively control the dot-gain and suitable to different output conditions.