...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Neural processing letters >Multitask Classification Method Based on Label Correction for Breast Tumor Ultrasound Images
【24h】

Multitask Classification Method Based on Label Correction for Breast Tumor Ultrasound Images

机译：基于乳腺肿瘤超声图像标签校正的多任务分类方法

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

To enable deep learning-based computer-aided diagnosis to achieve excellent performance in differentiating benign and malignant breast tumors in ultrasound images, a large number of labeled training samples must be collected. However, it is difficult to acquire sufficient samples due to the high costs of data collection and labeling. Fortunately, breast ultrasound images have two labels from different sources of domain knowledge: the biopsy results are "clean" labels, and the Breast Imaging Reporting and Data System (BI-RADS) score functions as a "noisy" label. Based on these two label types, we propose a multitask classification method based on label distribution correction (MTLC-Net). In our method, we propose different tasks to address the noisy and clean labels. Specifically, we propose a label distribution correction task for noisy labels that includes jointly training the network parameters and soft labels. The model is generalizable and robust by correcting the noisy label distribution based on the BI-RADS score, and it extracts knowledge from the noisy label task to improve the learning in the clean-label task. We conducted extensive comparisons with existing methods. Our method achieved a classification accuracy of 75.8%, a precision of 73.0%, a recall of 80.1% and an F1 score of 0.764-results that are significantly better than those of the existing state-of-the-art methods for differentiating benign and malignant breast tumors in ultrasound images.
机译：为了实现基于深度学习的计算机辅助诊断，在超声图像中差异化良性和恶性乳腺肿瘤来实现优异的性能，必须收集大量标记的训练样品。然而，由于数据收集和标记的高成本，难以获得足够的样品。幸运的是，乳房超声图像具有来自不同领域知识来源的两个标签：活组织检查结果是“清洁”标签，乳房成像报告和数据系统（BI-RADS）得分作为“嘈杂”标签。基于这两个标签类型，我们提出了一种基于标签分布校正（MTLC-NET）的多任务分类方法。在我们的方法中，我们提出了不同的任务来解决嘈杂和干净的标签。具体而言，我们为嘈杂标签提出了一个标签分布校正任务，包括联合培训网络参数和软标签。通过基于BI-RADS分数校正嘈杂的标签分布，该模型是概括和强大的，并且它从嘈杂的标签任务中提取知识，以改善清洁标签任务中的学习。我们与现有方法进行了广泛的比较。我们的方法实现了75.8％的分类精度，精度为73.0％，召回量为80.1％，F1得分为0.764分，结果明显优于现有的良性方法的现有技术方法。超声图像中的恶性乳腺肿瘤。

著录项

  • 来源
    《Neural processing letters》 |2021年第2期|1453-1468|共16页
  • 作者

    Cao Zhantao; Yang Guowu; Li Xiaoyu; Chen Qin; Wu Jinzhao;

  • 作者单位

    Univ Elect Sci & Technol China Sch Comp Sci & Engn Chengdu 611731 Peoples R China;

    Univ Elect Sci & Technol China Sch Comp Sci & Engn Chengdu 611731 Peoples R China;

    Univ Elect Sci & Technol China Sch Informat & Software Engn Chengdu 611731 Peoples R China;

    Univ Elect Sci & Technol China Sichuan Prov Peoples Hosp Sch Med Chengdu 610072 Peoples R China;

    GuangXi Univ Nationalities Coll Math & Phys Nanning 530006 Peoples R China;

  • 收录信息
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

    Deep learning; Multitask learning; Breast ultrasound image; Noisy label;

    机译：深入学习;多任务学习;乳房超声图像;嘈杂的标签;

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号