Painting is done according to the artist’s style.The most representative of the style is the texture and shape of the brush stroke.Computer simulations allow the artist’s painting to be produced by taking this strok...Painting is done according to the artist’s style.The most representative of the style is the texture and shape of the brush stroke.Computer simulations allow the artist’s painting to be produced by taking this stroke and pasting it onto the image.This is called stroke-based rendering.The quality of the result depends on the number or quality of this stroke,since the stroke is taken to create the image.It is not easy to render using a large amount of information,as there is a limit to having a stroke scanned.In this work,we intend to produce rendering results using mass data that produces large amounts of strokes by expanding existing strokes through warping.Through this,we have produced results that have higher quality than conventional studies.Finally,we also compare the correlation between the amount of data and the results.展开更多
In computer graphics, non-photorealistic rendering(NPR) is an important branch. NPR technology is to achieve a variety of artistic effects through the computer, such as oil painting, cartoon, watercolor and other effe...In computer graphics, non-photorealistic rendering(NPR) is an important branch. NPR technology is to achieve a variety of artistic effects through the computer, such as oil painting, cartoon, watercolor and other effects. The purpose of this paper is to automatically convert the 3D model into two-dimensional Chinese ink painting effect on graphics processing unit(GPU), and has improved the traditional algorithm that has some lacks of rendering effect. The algorithm is divided into two major steps: contour line rendering, and interior rendering. For contour line rendering, on the basis of the traditional extraction of contours, this paper adds self-bold, particle diffusion and other steps. For interior rendering, on the basis of the traditional layered rendering of diffuse lighting, two methods are added. The first method is based on the mean filtering, and the filter kernel is stratified by the principle of percentage-closer soft shadows(PCSS). The second method is Noise texture mapping, to obtain an ink diffusion effect.展开更多
Recently,there has been an upsurge of activity in image-based non-photorealistic rendering(NPR),and in particular portrait image stylisation,due to the advent of neural style transfer(NST).However,the state of perform...Recently,there has been an upsurge of activity in image-based non-photorealistic rendering(NPR),and in particular portrait image stylisation,due to the advent of neural style transfer(NST).However,the state of performance evaluation in this field is poor,especially compared to the norms in the computer vision and machine learning communities.Unfortunately,the task of evaluating image stylisation is thus far not well defined,since it involves subjective,perceptual,and aesthetic aspects.To make progress towards a solution,this paper proposes a new structured,threelevel,benchmark dataset for the evaluation of stylised portrait images.Rigorous criteria were used for its construction,and its consistency was validated by user studies.Moreover,a new methodology has been developed for evaluating portrait stylisation algorithms,which makes use of the different benchmark levels as well as annotations provided by user studies regarding the characteristics of the faces.We perform evaluation for a wide variety of image stylisation methods(both portrait-specific and general purpose,and also both traditional NPR approaches and NST)using the new benchmark dataset.展开更多
提出了一种新的基于线性卷积积分(Line Integral Convolution)自动铅笔画生成方法。提出的方法改进了已有的铅笔画生成方法,首先利用基于图的图像分割方法实现快速有效的区域分割,其次提出一种新的基于区域的白噪声和纹理方向生成方法...提出了一种新的基于线性卷积积分(Line Integral Convolution)自动铅笔画生成方法。提出的方法改进了已有的铅笔画生成方法,首先利用基于图的图像分割方法实现快速有效的区域分割,其次提出一种新的基于区域的白噪声和纹理方向生成方法。实验表明提出的方法更接近于真实的铅笔画效果。展开更多
基金This research was supported by the Chung-Ang University Research Scholarship Grants in 2017.
文摘Painting is done according to the artist’s style.The most representative of the style is the texture and shape of the brush stroke.Computer simulations allow the artist’s painting to be produced by taking this stroke and pasting it onto the image.This is called stroke-based rendering.The quality of the result depends on the number or quality of this stroke,since the stroke is taken to create the image.It is not easy to render using a large amount of information,as there is a limit to having a stroke scanned.In this work,we intend to produce rendering results using mass data that produces large amounts of strokes by expanding existing strokes through warping.Through this,we have produced results that have higher quality than conventional studies.Finally,we also compare the correlation between the amount of data and the results.
基金Supported by National Natural Science Foundation of China(NSFC)(61672260)
文摘In computer graphics, non-photorealistic rendering(NPR) is an important branch. NPR technology is to achieve a variety of artistic effects through the computer, such as oil painting, cartoon, watercolor and other effects. The purpose of this paper is to automatically convert the 3D model into two-dimensional Chinese ink painting effect on graphics processing unit(GPU), and has improved the traditional algorithm that has some lacks of rendering effect. The algorithm is divided into two major steps: contour line rendering, and interior rendering. For contour line rendering, on the basis of the traditional extraction of contours, this paper adds self-bold, particle diffusion and other steps. For interior rendering, on the basis of the traditional layered rendering of diffuse lighting, two methods are added. The first method is based on the mean filtering, and the filter kernel is stratified by the principle of percentage-closer soft shadows(PCSS). The second method is Noise texture mapping, to obtain an ink diffusion effect.
文摘Recently,there has been an upsurge of activity in image-based non-photorealistic rendering(NPR),and in particular portrait image stylisation,due to the advent of neural style transfer(NST).However,the state of performance evaluation in this field is poor,especially compared to the norms in the computer vision and machine learning communities.Unfortunately,the task of evaluating image stylisation is thus far not well defined,since it involves subjective,perceptual,and aesthetic aspects.To make progress towards a solution,this paper proposes a new structured,threelevel,benchmark dataset for the evaluation of stylised portrait images.Rigorous criteria were used for its construction,and its consistency was validated by user studies.Moreover,a new methodology has been developed for evaluating portrait stylisation algorithms,which makes use of the different benchmark levels as well as annotations provided by user studies regarding the characteristics of the faces.We perform evaluation for a wide variety of image stylisation methods(both portrait-specific and general purpose,and also both traditional NPR approaches and NST)using the new benchmark dataset.
