Perceiving stroke information from color-blindness images

Dichromats (a human with red-green color blindness considered in this paper) cannot always perceive meaningful visual information due to their deficiencies in eye cells. Stroke information extraction from a color-blindness image (CBI) is presented in order to deliver direct and effective information to dichromats. A CBI is first transformed into the pattern-highlighted image by means of color component analysis, pattern attention as well as thresholding. Then the maximal effective information in the highlighted image is computed to form a gray image. Based on the assistance of gray thinning algorithm, the useful stroke information in CBI is obtained. Experimental results on the often-used Ishihara test plates confirm that the perceived stroke information agrees with the normal perception.