基于自适应局部非线性回归的颜色校正算法  

Color correction method based on self-adaptive local nonlinear regression

作　　者：马玫娟[1,2] 蔡轶珩[1,2] 张新峰[1] 徐晓昭[1] 

机构地区：[1]北京工业大学信号与信息处理研究室,北京100124 [2]北京工业大学机械工程与应用电子技术学院,北京100124

出　　处：《计算机工程与应用》2010年第13期164-167,共4页Computer Engineering and Applications

基　　金：国家自然科学基金重大项目No.60431020;国家自然科学基金No.60431020;No.60472036;北京市自然科学基金No.4092009~~

摘　　要：颜色是图像的重要信息。许多颜色校正算法都采用精度较高的查找表法。为了更好地拟合颜色空间之间复杂的映射关系,在自适应局部线性回归颜色校正模型的基础上提出了基于自适应局部非线性回归的颜色校正模型,在小样本情况下,自适应地选择插值点的个数,利用局部非线性回归模型优化权值,建立三维的查找表,实现较好的颜色校正效果。实验证明基于自适应局部非线性回归的颜色校正模型的校正精度整体高于基于自适应局部线性回归的颜色校正模型的校正精度。Color is important information of an image.Several color correction methods are used to establish Look-up table,which is one of popular ways for color reproduction,so that high accuracy can be achieved.A color correction method based on selfadaptive local nonlinear regression is presented,for nonlinear regression can better fit the complex mapping of color spaces.Under small sampling condition,the color correction method adaptively selects the most optimized neighborhood size and applies local nonlinear regression to optimize the weights,then establishes 3D Look-up table so that ideal performance can be achieved.The results of experiment show that the accuracy of color correction method based on self-adaptive local nonlinear regression is higher than that of the color correction model based on self-adaptive local linear regression.

关 键 词：局部非线性回归 查找表 颜色校正 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]