摘要
针对传统的图像彩色化算法易产生的图像边缘颜色模糊等问题,提出一种基于变分法和偏微分方程的图像彩色化算法,结合整体变分(TV)模型和金兹堡-朗道(GL)模型的特点,通过求泛函极值建立偏微分方程,使图像的光滑区域和边界区域都能有效的实现颜色扩散,对偏微分方程进行数值求解得到彩色化结果。实验结果表明,与其它偏微分方程彩色化技术相比,该方法能得到清晰自然的彩色化图像和较高的图像质量。
A gray image colorization algorithm based on improved total variation(TV) and partial differential equation(PDE) is proposed to overcome problems of color blurring near edges in some colorization algorithms.According to the character of TV and Ginzburg-Landau(GL),a partial differential equations is established by evaluating the minimization of functional.The colors can be diffusion effectively in the smooth and boundary zone of gray image.A colorized image is obtained by numerically solving the PDE.Compared with some other PDE methods,the proposed method can produce colorized images with better visual quality.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第9期3531-3534,共4页
Computer Engineering and Design
基金
上海应用技术学院科研基金项目(KJ2009-15)
关键词
图像彩色化
整体变分
金兹堡-朗道
泛函
偏微分方程
image colorization
total variation
Ginzburg-Landau
functional
partial differential equations