《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (9): 2955-2962.DOI: 10.11772/j.issn.1001-9081.2022081159

• 多媒体计算与计算机仿真 • 上一篇    下一篇

基于EfficientNetV2和物体上下文表示的胃癌图像分割方法

周迪1,2, 张自力1,2(), 陈佳1,3, 胡新荣2,3, 何儒汉2,3, 张俊4   

  1. 1.武汉纺织大学 计算机与人工智能学院, 武汉 430200
    2.武汉纺织大学 湖北省服装信息化工程技术研究中心, 武汉 430200
    3.武汉纺织大学 纺织服装智能化湖北省工程研究中心, 武汉 430200
    4.武汉工程大学 计算机科学与工程学院, 武汉 430205
  • 收稿日期:2022-08-07 修回日期:2022-11-03 接受日期:2022-11-14 发布日期:2023-01-11 出版日期:2023-09-10
  • 通讯作者: 张自力
  • 作者简介:周迪(1997—),男,湖北武汉人,硕士研究生,CCF会员,主要研究方向:机器学习、图像处理
    陈佳(1982—),女,湖北武汉人,讲师,博士,CCF会员,主要研究方向:图像处理、模式识别
    胡新荣(1973—),女,湖北武汉人,教授,博士,CCF会员,主要研究方向:虚拟现实、机器学习
    何儒汉(1974—),男,湖北宜昌人,教授,博士,CCF会员,主要研究方向:人工智能、计算机视觉;
    张俊(1975—),男,湖北随州人,教授,博士,CCF会员,主要研究方向:机器学习、智能制造
  • 基金资助:
    湖北省教育厅科学技术研究计划项目(B2017066)

Stomach cancer image segmentation method based on EfficientNetV2 and object-contextual representation

Di ZHOU1,2, Zili ZHANG1,2(), Jia CHEN1,3, Xinrong HU2,3, Ruhan HE2,3, Jun ZHANG4   

  1. 1.School of Computer Science and Artificial Intelligence,Wuhan Textile University,Wuhan Hubei 430200,China
    2.Engineering Research Center of Hubei Province for Clothing Information,Wuhan Textile University,Wuhan Hubei 430200,China
    3.Hubei Provincial Engineering Research Center for Intelligent Textile and Fashion,Wuhan Textile University,Wuhan Hubei 430200,China
    4.School of Computer Science and Engineering,Wuhan Institute of Technology,Wuhan Hubei 430205,China
  • Received:2022-08-07 Revised:2022-11-03 Accepted:2022-11-14 Online:2023-01-11 Published:2023-09-10
  • Contact: Zili ZHANG
  • About author:ZHOU Di, born in 1997, M. S. candidate, His research interests include machine learning, image processing.
    CHEN Jia, born in 1982, Ph. D., lecturer. Her research interests include image processing, pattern recognition.
    HU Xinrong, born in 1973, Ph. D., professor. Her research interests include virtual reality, machine learning.
    HE Ruhan, born in 1974, Ph. D., professor. His research interests include artificial intelligence, computer vision.
    ZHANG Jun, born in 1975, Ph. D., professor. His research interests include machine learning, intelligent manufacturing.
  • Supported by:
    Science and Technology Research Program of Department of Education of Hubei Province(B2017066)

摘要:

针对U-Net上采样过程容易丢失细节信息,以及胃癌病理图像数据集普遍偏小,容易出现过拟合的问题,提出一种基于改进U-Net的自动分割胃癌病理图像模型EOU-Net。EOU-Net在U-Net模型的基础上,将EfficientNetV2作为骨干特征提取网络,以增强网络编码器的特征提取能力。在解码阶段,基于物体上下文表示(OCR)探究细胞像素间的关系,并使用改进后的OCR模块解决上采样图像的细节丢失问题;然后,使用验证阶段增强(TTA)后处理对输入图像进行翻转和不同角度旋转后分别预测这些图像,再通过特征融合的方式将多个输入图像预测结果进行合并,进一步优化网络的输出结果,从而有效解决医学数据集较小的问题。在SEED、BOT以及PASCAL VOC 2012数据集上的实验结果表明,与OCRNet相比,EOU-Net的平均交并比(MIoU)分别提高了1.8、0.6以及4.5个百分点。可见EOU-Net能得到更准确的胃癌图像分割结果

关键词: 语义分割, U-Net, EfficientNetV2, 物体上下文表示, 胃癌

Abstract:

In view of the problems that the upsampling process of U-Net is easy to lose details, and the datasets of stomach cancer pathological image are generally small, which tends to lead to over-fitting, an automatic segmentation model for pathological images of stomach cancer based on improved U-Net was proposed, namely EOU-Net. In EOU-Net, based on the existing U-Net model, EfficientNetV2 was used as the backbone, thereby enhancing the feature extraction ability of the network encoder. In the decoding stage, the relations between cell pixels were explored on the basis of Object-Contextual Representation (OCR), and the improved OCR module was used to solve the loss problem of the upsampled image details. Then, the post-processing of Test Time Augmentation (TTA) was used to predict the images obtained by rollover and rotations at different angles of the input image respectively, and then the prediction results of these images were combined by feature fusion to further optimize the output results of the network, thereby solving the problem of small medical datasets effectively. Experimental results on datasets SEED, BOT and PASCAL VOC 2012 show that the Mean Intersection over Union (MIoU) of EOU-Net is improved by 1.8, 0.6 and 4.5 percentage points respectively compared with that of OCRNet. It can be seen that EOU-Net can obtain more accurate segmentation results of stomach cancer images.

Key words: semantic segmentation, U-Net, EfficientNetV2, Object-Contextual Representation (OCR), stomach cancer

中图分类号: