Brush-based Non-photorealistic Drawing

Non-photorealistic rendering is one of the most active research topics fields of computer graphics. In this paper, based on the research of SBR classic algorithm, The author has studied the color transfer algorithm,the multilayer adaptive algorithm of SBR, painting style of pencil drawing, and other related algorithms.In this paper, first, we analyze the principle of SBR and discuss the principle of SBR discussed the optimization algorithm and the greedy algorithm in detail. Next, given the two images, namely color image and shape image, authors first transform their RGB color space to (?) aβcolor space whose three principal axes are orthogonal to decorrelate the color space of them, based on an adaptively and locally sampling scheme, RGB to (?)αβspace transformation and statistical method is applied to acquire the color from color image and the shape from shape information image, then synthesize them. Then, we study the multilayer adaptive algorithm of SBR, the main idea is to simulate the process of painter, using source image's gradient field can control strokes' movement, drawing a layer by layer on the canvas with the brush size from big to small , the algorithm designed by a different brush shape and parameters to produce a different artistic style. Finally, we study an effective algorithm of painting style of pencil drawing, the main idea is to use LIC algorithm to simulate pencil texture, use objects gradient information to get the outline edge in the image,and then synthesize the verge and texture to generate pencil drawing style images.