Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

Isolation, purification, and antioxidant activities of degraded polysaccharides from Enteromorpha prolifera.

International journal of biological macromolecules·2015
Same author

Bootstrap percolation on spatial networks.

Scientific reports·2015
Same author

Mechanistic Understanding of the Divergent Cyclizations of o-Alkynylbenzaldehyde Acetals and Thioacetals Catalyzed by Metal Halides.

Chemistry (Weinheim an der Bergstrasse, Germany)·2015
Same author

Quantification of the Iodine Content of Perigastric Adipose Tissue by Dual-Energy CT: A Novel Method for Preoperative Diagnosis of T4-Stage Gastric Cancer.

PloS one·2015
Same author

[Anterior interhemispheric approach through the lamina terminalis for large intra- and extra-ventricular craniopharyngiomas].

Zhonghua wai ke za zhi [Chinese journal of surgery]·2015
Same author

Syringe Injectable Electronics: Precise Targeted Delivery with Quantitative Input/Output Connectivity.

Nano letters·2015

相关实验视频

Updated: Jun 17, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

380

域互动对比学习和原型导向的自我训练,用于跨域的多重体细分.

Ziru Lu, Yizhe Zhang, Yi Zhou

    IEEE transactions on medical imaging
    |August 14, 2024
    PubMed
    概括

    这项研究引入了在结肠镜图像中聚细分的新框架,提高了不同设备的准确性. 域互动对比学习和原型导向自训练 (DCL-PS) 方法提高了模型在未见数据上的性能.

    科学领域:

    • 医疗成像医学成像
    • 人工智能的人工智能
    • 计算机视觉 计算机视觉

    背景情况:

    • 在结肠镜检查中精确的聚细分对于结肠直肠癌的诊断和治疗至关重要.
    • 深度学习模型在来自不同成像设备的数据集上的性能退化中扎.
    • 无监督域调整 (UDA) 方法旨在使用标记的源数据和未标记的目标数据来弥合域差距.

    研究的目的:

    • 提出一个新的框架,域互动对比学习和原型导向自训练 (DCL-PS),用于跨域的多片细分.
    • 解决现有的UDA方法的局限性,包括忽视域智能表示和伪标签不确定性.

    主要方法:

    • 域互动对比学习 (DCL) 具有域混合原型更新策略,以区分跨域类明智的特征.
    • 基于对比学习的交叉一致性培训 (CL-CCT) 增强编码器特征提取.
    • 原型导向的自我训练 (PS) 具有动态像素加权,以提高伪标签质量.

    主要成果:

    • 拟议的DCL-PS框架在目标域数据集上的聚细分方面表现出卓越的性能.
    • 该方法有效地减少了域差距,并改善了模型通用化.
    • 实验结果验证了DCL和PS策略的有效性.

    更多相关视频

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
    04:48

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

    Published on: November 30, 2022

    2.7K
    Deep Learning-Based Segmentation of Cryo-Electron Tomograms
    10:25

    Deep Learning-Based Segmentation of Cryo-Electron Tomograms

    Published on: November 11, 2022

    8.6K

    相关实验视频

    Last Updated: Jun 17, 2025

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
    04:48

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

    Published on: July 5, 2024

    380
    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
    04:48

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

    Published on: November 30, 2022

    2.7K
    Deep Learning-Based Segmentation of Cryo-Electron Tomograms
    10:25

    Deep Learning-Based Segmentation of Cryo-Electron Tomograms

    Published on: November 11, 2022

    8.6K

    结论:

    • 该DCL-PS框架提供了一个强大的解决方案,用于在结肠镜图像中跨域聚细分.
    • 拟议的策略增强了特征歧视和伪标签可靠性.
    • 这项工作有助于在结直肠癌查中更准确,更可靠地检测聚.