| 摘要: |
| 以色度学理论和神经网络为基础，在综合比较数码相机颜色特征化方法之后，采用了基于BP神经网络的颜色特征化方法。利用MATLAB实现了从RGB到XYZ及Lab的颜色空间转换，比较了2种转换模型的精度，统计分析了该模型产生训练误差和测试误差的主要原因，验证了RGB→Lab转换算法是一种较优和可靠的方法。 |
| 关键词: BP神经网络 数码相机颜色特征化 色彩空间转换 |
| DOI： |
| 分类号:TS803; TS801.3 |
| 基金项目: |
|
| Color Characterization for Digital Camera Based on BP Neural Network |
|
JIANG Fei-fei, XU Lan-ping, ZHENG Li-yang
|
|
Shanghai University, Shanghai 200072, China
|
| Abstract: |
| Color characterization method for digital cameras based on BP neural network was proposed after comparison of color characterization methods using the color theory and neural network theory. Color space conversion from RGB to XYZ and Lab was realized by using MATLAB. The precision of the two models was compared and the main causes of training and test error were analyzed statistically. The results showed that RGB to Lab color space version is better than RGB to XYZ. |
| Key words: BP neural network color characterization for digital camera color space conversion |
渝公网安备 50010702501716号