Stippling technique based on color analysis

초록

This paper presents a new stippling technique that effectively represents color images similar to the stippling works of artists. First, the proposed method creates a reference image from combining a smoothed image and with an edge image of the source. The color density function, defined by analyzing the information from the HSV color model of the reference image, is applied to the automatic computation of the number, size, and position of the points. Then, we use the weighted Penrose-based importance sampling to make points to be distributed randomly but almost evenly spaced in accordance with the color density within few seconds. In addition, our color jittering can generate efficiently stippling drawings having various colors. For more artistic representation, the points arc transformed into the oval-shaped ones and their orientations are defined according to the direction field of color image. Our application areas include animation, digital art, and video processing.