摘要
提出了一种用多层神经网络对图像进行非线性分割的方法，讨论了所用多层神经网络的学习速度的改进与训练样本的选择方法。实验表明，该多层神经网络系统可用于实时图像分割。
A new image nonlinear segmentation method based on multilayer neural network is proposed in this paper. The technique of improving the learing speed and the method of selecting learing patterns are discussed. The experiment shows that proposed multilayer neural network can be used for real time image segmentation and the good results can be obtained.
出处
《光学学报》
EI
CAS
CSCD
北大核心
1997年第1期74-78,共5页
Acta Optica Sinica
关键词
多层
前馈
神经网络
图像处理
非线性图像分割
multilayer feedforward neural network, real time learning, image nonlinear segmentation.
