Research on image color intelligence optimization for visual effects
In order to improve the visual effect of images,this article studies intelligent optimization methods for visual effect ori-ented image colors and proposes an ACE optimization algorithm based on LLLUT.Firstly,study the standard automatic color balance(ACE)algorithm;Then,in response to the shortcomings of the standard ACE algorithm,improvements were made to reduce its com-putational complexity and improve the speed and performance of the algorithm in image equalization processing.The Local Linear Lookup Table(LLLUT)method was selected to optimize the ACE algorithm;Finally,experimental testing was conducted on the fea-sibility and effectiveness of the LLLUT based ACE optimization algorithm.The experimental results show that the average values of the corrected images using the LLLUT based ACE optimization algorithm significantly decrease and are adjusted between 120 and 130,which is in line with the optimal visual range of the human eye.Among them,for images with insufficient lighting,the mean of our algorithm has increased by 4.726 5,the standard deviation has increased by 2.407 3,and the entropy value has increased by 0.0842 compared to the standard ACE algorithm;For overexposed images,the standard deviation of our algorithm has increased by 3.1373 compared to the mean of the standard ACE algorithm,and the entropy value has increased by 0.164 2;For contrast haze images,the standard deviation of our algorithm has increased by 33.745 2 compared to the original image,and by 10.121 9 compared to the standard ACE algorithm.When performing image correction,for images with a size of 300*400,the standard ACE algorithm re-quires 701.223 seconds,while the algorithm in this article only requires 2.727 seconds;For images with a size of 1 600*1 600,the standard ACE algorithm requires 418 247 s,while the algorithm in this article only requires 336.414 s.In summary,it can be seen that the ACE optimization algorithm based on LLLUT proposed in this article is feasible and effective,which not only improves the visual effect of the image,but also effectively improves the computational speed of the algorithm.
万方数据
维普
