基于多个前向神经网络和遗传算法的边界检测法

Edge Detection Using Multiple Feedforward Neural Networks and Genetic Algorithms

摘要: 本文提出了一种基于前向神经网络和遗传算法的图象边界检测方法, 这种方法由多个三层子前向神经网络并置连接加一个MAXNET子神经网络形成, 每个前向神经子网络的隐含层神经元个数和所有的权重系数由遗传算法优化确定。实验结果表明, 本文提出的这种边界检测方法抗噪声能力强, 检测到的边界位置更为准确。

Abstract: A novel algorithm using neural networks and genetic algorighms to detect image edge is proposed in this paper.It is composed of multiple parallel feedforward neural networks and a MAXNET net.The structure and weights of these parallel neural networks are optimized by genetic algorithms.The experiments show that the proposed method obtains more precise edge maps and is robust to noise.