基于注意力机制和残差结构的细胞核分割方法
A cell nucleus segmentation method based on attention mechanism and residual structure
刘国华 1闫克丁 1张亮 1刘叶楠 1马杭 1张祺1
作者信息
- 1. 西安工业大学电子信息工程学院,陕西西安 710021
- 折叠
摘要
病理细胞核的精准分割是病理学诊断的基础,然而当前算法针对带有核分裂象的乳腺癌细胞核自动分割效果差强人意.本文针对当前细胞核分割算法展开分析研究,并提出了一种基于注意力机制和残差结构相结合的U型网络(U_net)用于解决因核分裂象和非核分裂象细胞形态特征十分接近造成细胞核分割精度不足的问题.通过实验表明本文提出算法的平均像素准确率(mean pixel accuracy,MPA)和Mean_dice指标系数分别为0.74和0.82.与原有算法相比,训练指标分别提升了 11％和9％,证明本文算法的可行性.
Abstract
Accurate segmentation of pathological cell nucleus is the basis of pathological diagnosis,howev-er,the current algorithm for automatic segmentation of mitotic breast cancer cell nucleus is poor.This paper analyze and study the current cell nucleus segmentation algorithm.And a U-shaped network(U_net)based on a combination of attention mechanism and residual structure is proposed to solve the prob-lem of insufficient accuracy of cell nucleus segmentation due to the close morphological characteristics of mitotic and non-mitotic cells.The experiments show that the average pixel accuracy(MPA)and Mean_dice index coefficient of the proposed algorithm are 0.74 and 0.82.Compared with the original algo-rithm,the training indexes are improved by 11％and 9％,which proves the feasibility of the algorithm in this article.
关键词
图像处理/核分裂象分割/残差结构/注意力机制/深度学习Key words
image processing/mitotic cell segmentation/residual structure/attention mechanism/deep learning引用本文复制引用
基金项目
国家自然科学基金(11804263)
陕西省教育厅专项科研计划项目(21JK0819)
出版年
2023
NETL
NSTL
维普
万方数据