基于平均明暗熵差的人脸增强算法

Face Enhancement Alogrithm Based on Average Bright and Dark Entropy Difference

针对人类视觉对图像亮度的响应特征,提出了一种使用粒子群优化方法对图像灰度的非线性转换函数参数进行优化,用以使图像的平均明暗信息熵之差的绝对值最小,从而使图像的明暗灰度分布更加均匀的人脸增强算法.实验结果表明,该方法在增强图像的基础上能有效地提高图像在高灰度区域的视觉效果,对图像在高灰度区对比度有所损失的现象进行弥补,同时也能有效地提高整幅图像的视觉层次.

Focused on the relation between human vision and image brightness,the face enhancement algorithm is presented,in which partical swarm optimization is used to optimize the parameter of nonlinear gray transition function about an image in order to make the absolute value about the average bright and dark entropy difference lowest.The method can also make bright and dark gray level distribution more equal.Experiment results show that this method can not only improve the visual effect in high gray area of an image,but also make up the loss of contrast in that of an image. At the same time,the processed image has a better visual hierarchy.