摘要
针对脉冲耦合神经网络(PCNN)模型主要应用于灰度图像处理的局限性,利用脉冲发生器将颜色信息引入模型作为输入,与灰度信息共同控制神经元的内部行为,控制等灰度值的不同颜色区域分期点火,实现彩色图像的精确分割.双输入PCNN模型实现了彩色图像的分割,同时保持了PCNN模型对噪声的鲁棒性,从简单的仿真图像和实际图像两方面验证了此分割方法的有效性.
To overcome the limitation of the pulse-coupled neural network (PCNN) for its application in processing of color images, the hue data were introduced into the basic PCNN model by adopting the signal generator to control the internal activities of cells. The pixels with the same grey level and different hue data were pulsed separately in the double-input PCNN. It is indicated that the double-input PCNN can achieve the segmentation of color images and is robust to noises. The experimental results on synthetic images and natural images verify the validity of the model.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2009年第3期58-62,共5页
Journal of Harbin Institute of Technology
基金
哈尔滨青年科技创新人才基金资助项目(2008RFQXS037)