西北师范大学学报（自然科学版）2017，Vol.53Issue(2)：46-51,111,7.DOI:10.16783/j.cnki.nwnuz.2017.02.009
基于改进3D-PCNN的中药材彩色显微图像分割
Segmentation of traditional Chinese medicinematerials color microscopic images based on improvedthree dimensional pulse coupled neural network
摘要
Abstract
In order to effectively extract the target information from traditional Chinese medicine(TCM) materials color microscopic images,an improved image segmentation method for TCM materials color microscopic images based on three-dimensional pulse coupled neural network(3D-PCNN) is proposed in this paper.Firstly,the traditional model is simplified and improved from the perspective of suitable processing TCM materials color microscopic images.Secondly,within RGB color space,the maximum information entropy was utilized to selectively arrange images,and the results can be treated as the improved model of input.The image segmentation is achieved.Finally,the best segmentation result is selected by using the largest comprehensive criterion,compared with maximum Shannon entropy,minimum cross entropy and color contrast.The experiment results show that the proposed method can accurately and automatically segment the TCM materials color microscopic images,and overcomes the defects of maximum Shannon entropy,minimum cross entropy and color contrast simultaneously.It is superior to the traditional methods.关键词
改进三维脉冲耦合神经网络/RGB颜色空间/中药材彩色显微图像分割/分割准则Key words
improved 3D-PCNN/RGB color space/TCM materials color microscopic image segmentation/segmentation criterion分类
信息技术与安全科学引用本文复制引用
马冬梅,李金凤,刘勍..基于改进3D-PCNN的中药材彩色显微图像分割[J].西北师范大学学报（自然科学版）,2017,53(2):46-51,111,7.基金项目
国家自然科学基金资助项目(61461046) （61461046）
