数码相机颜色特性化两种方法对比研究——BP神经网络和多项式转换法

近年来随着数字技术以及彩色数字图像输入输出设备的不断发展和广泛使用，颜色在不同的设备之间精确地传递或再现成为该领域的重要课题。文章以彩色数码相机特性化为研究内容．分别采用BP神经网络和多项式转换方法。尽管由实验结果方面看，用BP神经网络方法要优于多项式转换，但BP神经网络方法所需训练样本数量较多、训练时间长、不易实现，而多项式转换方法便于实际应用．因此多项式转换方法更利于数码相机特性化。

A Comparative Study of the Digital Cameras Colorimetric Characterization by Means of BP Neural Networks and Polynomial Transforms

With development and wide use of color imaging devices recent years, communication and reproduction of color information among different devices is becoming an important subject. In this paper two general techniques, BP artificial neural networks and polynomial transforms, are compared for their usefulness in characterizing color cameras. The neural is shown to give the better performance once the parameters of the models are optimized. Since BP neural networks can be difficult and time-consuming to train, it is concluded that polynomial transforms offer the better alternative for camera characterization.