基于耦合势能衍射检测的改进DCNON分割算法

Improved DCNON segmentation algorithm based on coupled potential diffract-detection

分析了动态耦合振荡神经网络(DCNON)的数学模型和工作原理,以及在图像分割应用中的不足;根据振荡神经网络的动力学特点,借鉴机械波传播模型,提出一种基于耦合势能衍射检测的改进DCNON图像分割算法,通过改进动态耦合振荡神经网络框架和改善算法结构来提高DCNON的运算效率。通过MATLAB对改进DCNON算法的仿真,对比原始分割算法,验证了改进算法有效地压缩了运算时间。

This paper analyzed mathematical model, working rules and disadvantages in image segmentation application of dynamically coupled neural oscillator network （DCNON）. Then according to dynamic characters of oscillator neural networks and mechanical wave propagation model, proposed an improved DCNON segmentation algorithm based on coupled potential diffract-detection, improved the computing efficiency of the DCNON by improving the dynamic coupled oscillation neural network framework and the algorithm structure. By MATLAB simulation for improving DCNON algorithm, compared to the original algorithm, the algorithm can effectively compress computational time.