红外与激光工程, 2003, 32 (4): 422, 网络出版: 2006-04-28
神经网络边缘拟合
Marginal fitting based on neural network
摘要
提出了一种基于神经网络学习特性和集合论的边缘分段算法(Merge-Split) ,优点在于它的分段间断点、自由度满足自适应性,能在学习过程中实现而不必事先规定;进一步提出神经网络拟合不连续的图像边缘的方法,在无边缘数学函数形式的情况下,充分利用其学习函数的任意逼近和自学习性进行拟合,实践证明该方法可以克服以往边缘拟合方法的不足.
Abstract
A merge-split algorithm based on study property of neural net and sum aggregate is presented. Its merit is that its merge is disconnected a little and freedom satisfies the adaptability, and it is realized during learning rather than stipulating in advance. Furthermore, the method of fitting edge of uncontinuous image with neural net is given. The method can carry on the fit at self-learning, using its learning function fully under the circumstances without the mathematical function form of edge. Some shortages of former marginal fit methods can be overcomed.
车国锋, 刘忠领, 李军伟, 朱振福, 舒金龙. 神经网络边缘拟合[J]. 红外与激光工程, 2003, 32(4): 422. 车国锋, 刘忠领, 李军伟, 朱振福, 舒金龙. Marginal fitting based on neural network[J]. Infrared and Laser Engineering, 2003, 32(4): 422.