摘要
为了实现人眼感知信息最大化,结合人眼视觉系统亮度阈值特性提出一种具有目标函数的图像增强算法.首先在全局灰度映射约束条件下给出感知图像的计算方法,并以感知图像与原图像之间的互信息作为目标函数;然后采用动态规划算法实现了全局最优的灰度合并策略,以合并图像中人眼无法分辨的灰度;最后通过调整合并后灰度的取值实现目标函数最大化.实验结果表明,该算法具有良好的图像增强效果,所提目标函数合理有效.
In order to achieve the maximum of human eye perceived image information, an image enhancement algorithm with objective function was proposed considering human visual system brightness threshold charac-teristic. Firstly, put forward a method to obtain perceived image under the global grayscale map constraints, and make the mutual information between perceived image and the original image as objective function; Sec-ondly, achieve a global optimum grayscale merger strategy by using dynamic programming algorithm to merge the grayscales which human eye could not distinguish; Finally, adjust the combined grayscale to maximize the objective function. Experimental results demonstrate the proposed algorithm has a good treatment effect, and the proposed objective function is reasonable and effective.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2015年第10期1900-1906,共7页
Journal of Computer-Aided Design & Computer Graphics
关键词
图像增强
平均交互信息
人眼感知亮度
临界可见偏差
动态规划
image enhancement
average mutual information
perceived brightness
just noticeable difference
dynamic programming
