| Digital halftoning is a technique used by binary display devices such as printers and monitors to create illusion of continuous tone. In order to mimic analog halftoning techniques, clustered-dot ordered dithering or Amplitude Modulated (AM) halftoning produces this illusion by varying the size of printed dots according to the gray level of an original continuous image. But this technique always weights relationship between spatial resolution and gray resolution, that is to say, an halftoned image with high spatial resolution must be at expense of low gray resolution. As an alternative to AM halftoning, Frequency Modulation (FM) or stochastic halftoning, which produces binary patterns with high quality by varying the intensity of dot distribution, maximize the space resolution without loss of gray resolution. Typical FM halftoning such as error diffusion and DBS generates high quality binary patterns, but has complex computation. Moreover, for hardcopy devices such as raster printers and inkjet printers (especially raster printers) do not display isolated pixels, the printed binary patterns produced by FM halftoning is significant darker than its original due to dot gain. In this paper, I investigate the generation of a novel kind of binary patterns that have many components of middle frequency and a few components of high frequency of DFT of the binary patterns so that can robust to the dot gain produced by printers and that have a good visual quality. In this paper, I will design a series of filters based on the Contrast Sensitivity Function (CSF) of Human Visual System (HVS) model to optimize the " seed " binary pattern and binary patterns for all gray levels and employ statistics technique to contain proper clumps and pleasing distribution. HP printers with 300dpi and 600dpi resolution and visual distance of 36 inches will be used in my experiments. In the end of the paper, I will compare the quality of halftoned image by other halftoning techniques.
|
PDF Full Text Request