| 摘要: |
| 通过运用MATLAB7.8进行了BP和RBF的数值仿真，并作出了二者的转换误差图，研究了采用BP和RBF神经网络方法进行色空间转换的精度差异问题。结果表明：BP神经网络由于本身收敛速度慢及训练无记忆性等缺陷，整体性能低于RBF神经网络，即用RBF来解决色彩管理中的色空间转换问题更符合要求。 |
| 关键词: 印刷 色空间 前馈神经网络（BP） 径向基神经网络（RBF） |
| DOI： |
| 分类号:TS801.8; TS801.3 |
| 基金项目: |
|
| Application of Neural Network in Color Conversion |
|
LI Li-xia, ZHANG Yi-xin
|
|
Jiangnan University, Wuxi 214122, China
|
| Abstract: |
| BP and RBF neural network was applied in color space conversion with MATLAB 7.8 and the relevant conversion error diagrams were drawn. The accuracies of color space conversion by BP and RBF neural network were compared. The results showed that RBF neural network has better accuracy than BP because BP has the bug of slow convergence speed and lack of memory; RBF neural network can better solve problems of color space conversion of color management. |
| Key words: printing color space BP RBF |
渝公网安备 50010702501716号