Minimization of Halftone Noise in FLAT Regions for Improved Print Quality

The work in this thesis proposes a novel algorithm for enhancing the quality of flat regions in printed color image documents. The algorithm is designed to identify the flat regions based on certain criteria and filter these regions to minimize the noise prior and post Halftoning so as to make the hard copy look visibly pleasing. Noise prior to halftone process is removed using a spatial Gaussian filter together with a Hamming window, concluded from results after implementing various filtering techniques. A clustered dithering is applied in each channel of the image as Halftoning process. Furthermore, to minimize the post halftone noise, the halftone structure of the image is manipulated according to the neighboring sub-cells in their respective channels. This is done to reduce the brightness variation (a cause for noise) between the neighboring subcells. Experimental results show that the proposed algorithm efficiently minimizes noise in flat regions of mirumal gradient change in color images