摘要
交叉皮层模型是一种简化的脉冲耦合神经网络(Pulse Coupled Neural Network,PCNN),该模型保留了PCNN的脉冲耦合、发放同步脉冲、变阈值、产生脉冲波等重要特性,但调节参数比PCNN少,被认为具有比PCNN更加良好的图像处理能力。针对ICM模型图像分割效果取决于各个参数的优化调节,提出了一种克隆选择算法和ICM模型的图像自动分割算法。该算法的克隆选择算子以自适应方式调节,以最大信息熵作为亲和度函数,速度快,全局搜索效率高,实现了图像的自动分割。仿真表明,该算法有效可行。
The Intersecting Cortical Model （ICM） is a simplified model of Pulse--Coupled Neural Network （PCNN） model. It maintains important properties of pulse coupled, synchronize pulsing activity, vary threshold and production of pulse wave, and its adjustable parame ters is less. So it is considered to be better the ability of Image processing compared with PCNN, The effect of image segmentation based on ICM depends on optimal regulation of its adjustable parameters. Based on clonal selection algorithm and ICM, an algorithm of image auto- matic segmentation is proposed in this paper. The clone operator of this algorithm is adaptive regulation and its affinity is maximum information entropy. Its computation speed and search efficiency are very high. The Algorithm has realized image automatic segmentation. The simulation results show the algorithm is effective and feasible.
出处
《计算机测量与控制》
CSCD
北大核心
2009年第9期1807-1809,共3页
Computer Measurement &Control
基金
国家自然科学基金(60575013)
航空科学基金(20070153005)
航空支撑科技基金(07C53007)
关键词
交叉皮层模型
脉冲耦合神经网络
图像分割
克隆选择算法
intersecting cortical mode （ICM）
pulse coupled neural network （PCNN）
image segmentation
clonal selection algorithm
