摘要
肾透明细胞癌病理图像中细胞核的形态和位置信息对肾癌的良恶性分级诊断具有重要意义,为提高肾透明细胞癌细胞核分割的质量,本研究提出了基于深度卷积神经网络的细胞核分割方法。首先,根据标定的病理图像中细胞核轮廓,构建细胞核分割样本集;然后,深度卷积神经网络通过隐式特征学习对细胞核分割模型进行训练,避免人为设计特征;最后,利用细胞核分割模型对病理图像进行逐像素分割。实验结果表明,深度卷积神经网络的细胞核分割算法在肾透明细胞癌细胞核分割的像素准确率高达90.33%,细胞核分割性能稳定,深度卷积神经网络强大的鲁棒性和适应性使得肾透明细胞癌细胞核自动分割具有可能。
The shape feature and location information of clear cell renal cell carcinoma, s nucleus is importantbenign and malignant renal cell carcinoma. To improve nuclear segmentation accuracy, nuclear segmentation neural networks was proposed. First, nuclear sample dataset was formed according to the nuclear contour labeled by pathologist. Then,deep convolution neural net^vork extracted implicit nuclear feature instead of artificial nuclear characteristic and tion model were trained. Finally, the nuclear segmentation model did nuclear segmentation by showthat the clear cell renal cell carcinoma’s nucleus segmentation algorithmof deep convolution neural network is as high as 90. .3% in the nucleus pixel accuracy and the nucleus segmentation performance is stable. The strong robustness and adaptability of deep convo-lution neural net^vork makes nuclear auto - segmentation possible.
出处
《生物医学工程研究》
北大核心
2017年第4期340-345,共6页
Journal Of Biomedical Engineering Research
基金
国家自然科学基金资助项目(61273259)
江苏省自然科学基金资助项目(BK20141482)
关键词
分割
卷积神经网络
细胞核
肾透明细胞癌
逐像素
Segmentation
Convolution neural network
Nuclei
Clear cell renal cell carcinoma
Pixel - wise
