摘要
脉冲耦合神经网络(Pulse Coupled Neural Network,PCNN)是一种新型神经网络模型,作为研究图像分割的常用方法,一直广受关注。针对目前大量文献关注PCNN模型仿真实现研究的情况,本文基于PCNN模型提出了将最小交叉熵分割算法在FPGA硬件平台上进行实现。相比于传统的PCNN软件实现以及最大信息熵分割算法实现的方案,本文提出的图像分割方案最佳分割精准度高,具有处理速度快,实时性强,图像分割效果好的优势,应用范围更广,因此该设计具有较高实际应用价值。
Pulse Coupled Neural Network(PCNN) is a new type of neural network model as a common method for image segmentation, which has received wide attention. As a large number of researches on simulation of the PCNN model, a minimum cross-entropy based on PCNN image segmentation algorithm is implemented on FPGA platform in this paper.Compared to the solutions of traditional software implementation based on PCNN and maximum entropy segmentation algorithm implementation, the solution in this paper has the advantages of higher accuracy of best image segmentation, high speed and excellent real-time performance. So, this design has broader market application prospect.
出处
《电子设计工程》
2015年第5期128-133,共6页
Electronic Design Engineering
关键词
PCNN
图像分割
最佳分割
最小交叉熵
FPGA
Pulse Coupled Neural Network(PCNN)
image segmentation
best segmentation
minimum cross-entropy
FPGA