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

Using community syndromic surveillance to anticipate enterovirus-related school class suspensions: A real-time LSTM approach.

International journal of infectious diseases : IJID : official publication of the International Society for Infectious Diseases·2026
Same author

Referenceless reduction of spin-echo echo-planar imaging distortion with generative displacement mapping.

Magnetic resonance in medicine·2025
Same author

Deep learning system for the differential diagnosis of oral mucosal lesions through clinical photographic imaging.

Journal of dental sciences·2025
Same author

Local excision for middle-low rectal cancer after neoadjuvant chemoradiation: A retrospective study from a single tertiary center.

World journal of gastrointestinal oncology·2024
Same author

Effects of astaxanthin on depressive and sleep symptoms: A narrative mini-review.

Heliyon·2023
Same author

Total neoadjuvant treatment for MRI-stratified high-risk rectal cancer: a single-center, single-arm, prospective Phase II trial (PKUCH-R02).

Gastroenterology report·2023

相关实验视频

Updated: May 10, 2025

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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

319

多模态多器官图像细分使用持续学习,提高对任务的重视度.

Ming-Long Wu1,2, Yi-Fan Peng2

  • 1Department of Computer Science and Information Engineering, National Cheng Kung University, Tainan, Taiwan.

Medical physics
|April 23, 2025
PubMed
概括

增强对任务的专注 (eHAT) 使深度神经网络能够进行持续学习,用于多模式,多器官医学图像细分. 与以前的方法相比,这种方法显著降低了遗忘率.

科学领域:

  • 人工智能的人工智能
  • 医学图像分析 医学图像分析
  • 机器学习 机器学习

背景情况:

  • 深度神经网络 (DNN) 可以通过持续学习模仿人类大脑功能,使他们能够连续学习多个任务.
  • 目前用于医疗图像细分的持续学习方法仅限于单模式图像和特定的解剖位置.

研究的目的:

  • 引入和评估一种先进的持续学习技术,增强对任务的重视 (eHAT),用于基于DNN的多模式和多器官细分.
  • 为了评估eHAT在复杂的医学成像场景中的性能.

主要方法:

  • 利用了四个公共数据集 (腰椎CT/MRI,心脏MRI,脑MRI) 来对脊椎体,右心室和脑瘤进行细分.
  • 在三任务和四任务模型上测试了eHAT,将其性能与最先进的持续学习方法进行比较.
  • 使用忘记率 (子系数和豪斯多夫距离的差异) 和向后转移 (BWT) 来量化多任务性能.

主要成果:

  • 与原始HAT (-18.13%至-3.59%) 相比,eHAT显示遗忘率大幅提高 (三项任务为-2.51%至-0.60%,四项任务为-2.54%至-1.59%).
  • 使用eHAT的四任务U-net模型在显著更少的参数 (半个通道) 中实现了可比性能.
  • 与HAT (-22%至-4%) 相比,eHAT表现出明显优异的BWT (-3%至0%),这表明知识转移更好.
关键词:
持续的学习,持续的学习.很努力地把注意力集中在任务上.图像细分 图像细分医学图像医学图像

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

2.6K
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

430

相关实验视频

Last Updated: May 10, 2025

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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

319
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

2.6K
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

430

结论:

  • eHAT是第一个成功实现使用单个DNN实现多模式,多器官细分的持续学习的方法.
  • 拟议的eHAT方法为医学成像的持续学习提供了更好的遗忘率和增强的知识传输能力.