针对非真实感抽象绘制技术中可能存在的抽象程度整体趋于一致、缺少色彩层叠感的问题,提出一种自动的、基于图像重要度的抽象艺术风格绘制方法。首先通过梯度和显著度加权计算得到图像的重要度,并将该重要度引入抽象艺术风格绘制;在绘制过程中,为了保留图像的局部结构和方向特征,采用各向异性滤波平滑结构张量场,并沿张量场局部曲线方向进行积分卷积计算;最后采用改进的局部光照明 Phong 模型对抽象结果进行渲染,有效地突出不同光照条件下的艺术表现力,增强了艺术效果的色彩层叠感。实验结果表明,该方法能对图像局部区域进行不同程度的抽象,更好地保留图像中的局部细节和结构特征,更加符合格式塔理论,并可以产生具有色彩层叠感的抽象艺术效果。
The problems for non-photorealistic rendering with current sketch rendering methods are the coinci-dent rendering degree and lack of hierarchical physical appearance of paintings. The paper presents an automatic method for simulating sketch paintings via significance map technology. Firstly, a significance map is built up in a weighted gradient and saliency of input image, yet guides sketch artistic rendering. Secondly, anisotropic filter technique is used to smooth tensor field, and integral convolution method is adopted to generate sketch paintings along the local curve direction of this tensor field. Finally, to highlight artistic appeal of sketch under different lighting condition, the painting is rendered by the improved Phong illumination model, which can enhance hier-archical physical appearance of sketch artistic works. Experimental results show that the proposed method can be used efficiently to produce special paintings with hierarchical appearance, and this method can carry out different level of sketch to the local region while preserves feature of the local details.