Current Issue Cover
ICSSM: 改进的彩色素描模拟方法

潘龙,纪庆革,陈靖(中山大学数据科学与计算机学院)

摘 要
目的 如今人们已经设计了很多模拟油画、水彩、水墨等风格的非真实感绘制方法,而能够生成彩色素描的算法还不是很多。针对这一课题,本文在前人工作的基础上,结合线积分卷积与双色调映射技术,改进了一种彩色素描模拟方法。方法 该方法首先基于k-means聚类对彩色图像进行分割,通过计算色彩差异性为每个区域指定两种基本色,并利用双色调映射技术计算每种颜色的密度。而后利用线积分卷积分别生成两个基本色层的素描纹理,并将两层纹理相融合来生成彩色纹理。与此同时,利用霓虹变换生成素描轮廓线。最后,将轮廓与彩色纹理相融合来得到彩色素描效果。结果 实验证明,本文的方法能够实现由彩色图像到彩铅画的自动、实时转化。结论 本文的方法从轮廓和纹理两个角度模拟了真实的彩铅绘画过程。基于k-means聚类的分割方法得到的结果能够更好地反应彩色图像的颜色分布特性。通过色彩差异性计算指定基本色的策略提高了该环节的效率,满足了实时性要求。由于粉笔、蜡笔等绘画风格的调色与彩铅画类似,本文将不同颜色层上下叠加的方式可以扩展到对其它介质绘画的模拟当中。
关键词
ICSSM: An improved color sketch simulation method

Long Pan,Qingge Ji,Jing Chen(School of Data and Computer Science,Sun Yat-sen University,Guangzhou,510006)

Abstract
Objective Non-photorealistic Rendering is the technique which is committed to generate art style works, its goal is not to represent the authenticity of graphics, but to simulate their art features, even defects of the art. This concept was first proposed in the 1980s. After years of development, Non-photorealistic Rendering techniques are able to simulate a lot of painting styles, including oil paintings, watercolor paintings, Chinese ink paintings, pencil sketches, cartoons and so on. The sketch is a kind of painting which presents the light and shade of an object by lines. The color sketch is the sketch painted by a small number of color pencils. Except for the light and shade, the color sketch represents color features of an object as well. When painting a color sketch, people follows the way of the monochrome sketch, and renders different regions with disparate color pencils. As the types of pencils are quite limited, people usually need to blend two or more colors to obtain a desired color. Besides, the color mixing of sketches is different from that of oil paintings or watercolor paintings. Typically people mixes pigments on a palette, but when doing a color sketch, people draws disparate color layers directly on the paper, then the layers of stroke for each color are superimposed and interspersed to build the target color through optical blending. This feature creates a very special kind of texture and color style for the color sketch. Although many people are fond of sketches, it is however, not easy for everybody to master the skill of pencil drawing. On the basis of previous works and characteristics of color sketch, combining the duotone color reproduction with LIC techniques, we realized an improved way of generating color sketch. Method Firstly we fulfilled a monochrome sketch simulation process based on LIC. On this basis, by researching on the features of the color sketch, and combining with the duotones color reproduction technique, we realized a new approach for the color sketch generation based on color customization, which can automatically convert a digital image into a colored pencil drawing. The first step of this method is image segmentation, then we use a primary color as well as a secondary color to paint each segmented region. For each particular region, our method can calculate its color set automatically. Such kind of color customization can successfully simulate the color mixing of sketch paintings, and it can also be expanded to the simulation of similar painting styles, like the chalk drawing, the pastel drawing, etc. Besides, we propose a way to superpose the real paper texture into digital images, which will make the final effect looked closer to real paintings. As the profile is one of the most important elements, we achieve a more ideal way to generate better profile for the color sketch based on neon-transformation. To improve the contrast ratio, we zoom the color of the original image before generating the white noise, which generates a clearer hierarchy for the final effects. We also fulfill a fresh way for image segmentation based on K-means cluster, which obtains more suitable segmentation results for the color sketch generation. Result To validate the effectiveness of the proposed approach, we have conducted extensive experiments. Experimental results demonstrate that our approach outperforms several state-of-the-art approaches on challenging automatic and real-time transformation from color image to color painting, and bridging the gap between them. Comparing to Yamamoto’s method, we incorporate the idea of hierarchical method and color scale transformation into linear integral convolution with two-tone color sketch mapping technology. In this way we obtain sketch texture closer to hand-painted effect. Masui’s method can generate more soft texture. But the whole strokes fusion process is really complicated and time-consuming. Especially, it does not take the sketch contours into account. This could be error-prone to lose structure characteristics of the object in drawing effect. As for Kim’s method, it extracts basic image feature for sketch generation. It has high efficiency and can be well applied to mobile devices. However, it considers the contour perspective only and thus ignores the effect of sketch texture. That often suffers a drop of accuracy comparing to real color sketch. By contrast of these two approaches, the proposed method takes both contour and texture into consideration. Simulating color image from both contour and texture aspects helps in issuing object characteristics preservation and prevent potential unrealistic sketch transformation. And Way’s method can generate more soft and saturate but cluttering texture, thus missing the natural line sense of color sketch. To tackle with this, we calculate texture direction by transforming image space to frequency domain for energy statistical characteristics. The simulation results obtain obvious intensive line sense. In color fusion aspect, Way’s method is based on the color of the original image and just overlay strokes texture on the basic color. However, in real drawing process people can only use limited colors. To dispose this challenge, we have established the library of basic colors. All colors used for sketch simulating are fused from the library and by two basic color layers. Simulation results show robustness to changing imaging conditions as well as satisfy human visual sense. Conclusion In general, our study reveals new insights into drawing scheme, implementing method, algorithm complexity and interactivity when simulating color sketch. Firstly, the proposed method is accomplished using image transformation approach. Therefore, it can transform color image to color sketch or any other basic color specified by user interactively. Secondly, our study encompasses both contour and texture simulation. It facilitates for simulating a more realistic painting process. Thirdly, the proposed method finds inspiration from human drawing experience. We overlay and fuse different color layers to generate color texture. This way can also be employed to simulate a wide spectrum of painting style, such as chalk drawing and crayon drawing.
Keywords
QQ在线


关注微信