基于GRNN的RGB到L＊a＊b＊色彩空间的转换模型

Research of conversion model from RGB to L＊a＊b ＊ color space based on GRNN

针对目前色彩空间转换精度低的问题,建立了一种基于广义回归神经网络(GRNN)的色彩空间转换模型.将RGB色彩空间进行分割,并将RGB值用Adobe Photoshop软件中的拾色器进行手工转换,得到对应的L*a*b*值,从而得到建模样本和检验样本数据.用MATLAB 7.5软件编程,建立RGB到L*a*b*色彩空间的转换模型,并对其转换精度进行评价.检验结果表明,该转换模型精度高且训练速度极快,可应用于色彩管理的相关技术领域.

A color spaces conversion model based on general regression neural network （GRNN） was established in view of low conversion precision in other researches. Firstly, dividing up the RGB color space and converting RGB value manually by using the color picker of Adobe Photoshop software, the data of modeling samples and checking samples were obtained. Secondly,the conversion model from RGB to L＊a＊b＊color space was established through programming with MATLAB 7. 5 software. Finally, the conversion precision was evaluated. The checking result shows that the conversion model has a high precision, the training speed is very fast,and it can be used for relative technique fields of color management.