基于直方图的脉冲耦合神经网络(PCNN)多门限图像分割方法

参考了Eckhom等人近几年提出的脉冲耦合神经网络(Pulse Coupled Neuml Network)模型，结合灰度直方图分割技术，提出了一种新的多门限图像分割方法-基于直方图的脉冲耦合神经网络(PCNN)，PCNN模型具有弥补时隙和空隙的特点，因此应用这一网络模型进行图像分割，可以得到较完整的区域边角信息，从而达到理想的分割效果．经实验证明，本文提出的方法较原来的PCNN网络，运算速度和分割效果都有了很大提高。

PULSE COUPLED NEURAL NETWORKS BASED ON HISTOGRAM FOR IMAGE MULTI-THRESHOLD SEGMENTATION

This paper presents a new multi - threshold segmentation method - pulse coupled neural network based on histogram. The algorithm is based on pulse coupled neural network (PCNN) presented firstly by Eckhom and the conventional histogram method. The algorithm of PCNN has some virtues for image segmentation, such as covering time intervals and space gaps. Thus it can achieve prefect result to segment image through PCNN, as the result holds more information of corners and edges. The experiments prove that our algorithm improves remarkably the speed and result of image segmentation.