Lin presents new dithering algorithm at IMAGAPP 2011
Derfen Lin, a PhD candidate advised by Prof. Jie Wang, presented her paper, “Producing Automated Mosaic Art Images of High Quality with Restricted and Limited Color Palettes,” at the IMAGAPP conference, held in Algarve, Portugal, March 5-7, 2011.
IMAGAPP is part of VISIGRAPP, a joint international conference on computer vision, imaging, and computer graphics theory and applications.
Lin studied the problem of mosaic art images, which are made from bricks, tiles, or counted cross-stitch patterns.
Artists need to divide the original image into small parts of reasonable sizes and shapes, and represent the colors of each part using just one closest color selected from a given color palette, a process called “dithering.”
Using standard methods to automate this process, the resulting mosaic image may contain undesirable visual artifacts of patches and color bandings.
In her study, Lin presented a new error-diffusion scheme, called Four-Way Block dithering (FWB), which corrects certain artifacts caused by existing dithering methods.
Lin showed that FWB can better retain the original structure and reduce unstructured artifacts. She also showed that FWB dithering produces much better peak signal-to-noise ratios on mosaic images over those generated by previous methods (please see image below for an example).
A copy of her paper is available at http://www.cs.uml.edu/~dlin/derfen_files/IMAGAPP_Derfen.pdf.
Image (a) shows dithering with previous approach; it is clearly noticeable that Mona Lisa’s right eye is blurry and looks half-closed. Image (b) shows use of the FWB algorithm. The eye problem is corrected.