A method for extracting handwritten onomatopoeia characters from comic images was developed based on the color feature and the width feature of characters. The segmentation using Delaunay triangulation and the color clustering with fuzzy c-means algorithm are applied. The effectiveness of the method was confirmed.
