基于脉冲耦合神经网络的Live wire新方法的仿真研究

提出了一种基于脉冲耦合神经网络(PCNN)的Live wire分割新方法,Live wire方法是把图像当作一个连通图,在边上定义一个代价函数,然后通过搜索最短路径来找物体的边界,把用户指定的物体边界上的两点之间的最短路径当作物体的边界。最短路径是Live wire方法的重要步骤,本文在介绍脉冲耦合神经网络的基本工作原理和特性的基础上,将改进的脉冲耦合神经网络算法引入到Live wire边缘检测的算法中,用于最短路径算法的研究。并在改进算法中应用路径封锁和在线训练来提高算法的准确性和应用性。

New Algorithm of Live Wire Based on PCNN Simulation Research

An improved live wire method for image segmentation was proposed based on the pulse coupled neural network (PCNN). Live wire image as a graph method is the definition of the edge of a cost function. Then through finding the shortest path to search objects edge, user specified object boundary as the shortest path between two points on the object's edge. Shortest Path is the most important step of live wire. The basic principle and characteristics of coupled neural network algorithm, the improved PCNN was introduced to Live wire, for the Live wire of the shortest path algorithm. The operation and application of path cooling and on-the-fly methods were introduced. Experiments show that the algorithm can obtain the boundary of the desired objects and improve the accuracy and applicability.