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

Application of high-frequency ultrasound in the diagnosis of subungual glomus tumor: A single center retrospective study.

Journal of the American Academy of Dermatology·2026
Same author

Utility of High-Frequency Ultrasound in Preoperative Evaluation of the Thickness of Cutaneous Melanoma.

Diagnostics (Basel, Switzerland)·2026
Same author

Role of non-invasive serological and CT imaging parameters in assessing portal hypertensive gastropathy severity.

BMC gastroenterology·2025
Same author

Pre-activation of T cell immunity potentiates ferroptotic cell death through arachidonic acid hybridized nanovesicles.

Journal of nanobiotechnology·2025
Same author

High-Frequency Ultrasound Quantification of Endophytic Growth Pattern: A Potential Imaging Indicator for Grading Cutaneous Squamous Cell Carcinoma Differentiation.

Dermatology (Basel, Switzerland)·2025
Same author

Targeting tumor-associated CCR2<sup>+</sup> macrophages to inhibit pancreatic cancer recurrence following irreversible electroporation.

Science advances·2025

相关实验视频

Updated: Jul 4, 2025

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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

405

变压器使CNN更好地适用于基于脚的医疗图像细分.

Zihan Li, Yuan Zheng, Dandan Shan

    IEEE transactions on medical imaging
    |February 7, 2024
    PubMed
    概括

    ScribFormer是一个新的CNN-Transformer混合模型,通过使用有限的涂注释来增强医疗图像细分. 它有效地捕捉了当地和全球的特征,优于现有的方法,甚至是完全监督的方法.

    科学领域:

    • 医学图像分析 医学图像分析
    • 计算机视觉 计算机视觉
    • 人工智能的人工智能

    背景情况:

    • 具有编码器-解码器架构的卷积神经网络 (CNN) 框架是涂监督细分的标准.
    • 这些方法难以捕获全球形状信息,因为卷积层的局部受体场,限制了它们的有效性与稀疏的涂注释.

    研究的目的:

    • 介绍ScribFormer,一种新的CNN-Transformer混合模型,旨在改善涂监督的医疗图像细分.
    • 解决现有方法的局限性,从有限的草稿数据中学习全球形状上下文.

    主要方法:

    • 开发了ScribFormer,这是一个三分支架构,集成CNN,变压器和注意力引导类激活地图 (ACAM) 分支.
    • 融合了CNN的本地特色与变压器分支机构的全球代表.
    • 利用ACAM分支来统一浅层和深层卷积特征以提高性能.

    主要成果:

    • 在公共和私人数据集上,ScribFormer与最先进的涂监督细分方法相比,表现出更高的性能.
    • 提出的方法取得的结果与完全监督的细分技术相比或更好.
    • 混合方法有效地克服了纯粹基于CNN的方法在捕捉全球背景方面的局限性.

    结论:

    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

    2.8K
    Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
    06:48

    Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images

    Published on: January 7, 2019

    8.9K

    相关实验视频

    Last Updated: Jul 4, 2025

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

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

    Published on: July 5, 2024

    405
    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

    2.8K
    Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
    06:48

    Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images

    Published on: January 7, 2019

    8.9K
    • 通过有效地结合本地和全球功能学习,ScribFormer在涂监督的医疗图像细分方面取得了重大进展.
    • 该模型利用有限注释的能力超过了当前最先进的甚至完全监督的方法,突出了其实际应用的潜力.
    • 混合CNN-变压器架构为具有稀疏监督的具有挑战性的细分任务提供了强大的解决方案.