Misalignment between the color planes used to print color images creates undesirable artifacts in printed images. Color trapping is a technique used to diminish these artifacts. It consists of creating small overlaps between the color planes, either at the page description language level or the rasterized image level. Existing color trapping algorithms for rasterized images trap pixels independently. Once a pixel is trapped, the next pixel is processed without making use of the information already acquired. We propose a more efficient strategy which makes use of this information. Our strategy is based on the observation of some important properties of color edges. Combined with any existing algorithm for trapping rasterized images, this strategy significantly reduces its complexity. We implement this strategy in combination with a previously proposed color trapping algorithm (WBTA08). Our numerical tests indicate an average reduction of close to 38% in the combined number of multiplications, additions, and "if" statements required to trap a page, as compared with WBTA08 by itself.
feature extraction, image colour analysis, image registration
Date of this Version
Proc. SPIE - Int. Soc. Opt. Eng. (USA) 7528,(2010) 75280C (12 pp.)-75280C (12 pp.);