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

Late-Onset Filler-Induced Granuloma After Polycaprolactone-Based Filler Treated With High-Intensity Focused Ultrasound and Quantum Molecular Resonance Technology.

Dermatologic surgery : official publication for American Society for Dermatologic Surgery [et al.]·2022
Same author

Keloid of earlobe treated with a combination of pneumatic injector device and CO<sub>2</sub> fractional laser.

Dermatologic therapy·2022
Same author

Effect of A. polygama APEE (Actinidia polygama ethanol extract) or APWE (Actinidia polygama water extract) on wrinkle formation in UVB-irradiated hairless mice.

Journal of cosmetic dermatology·2022
Same author

Transforming growth factor‑β family and stem cell‑derived exosome therapeutic treatment in osteoarthritis (Review).

International journal of molecular medicine·2022
Same author

Safety and efficacy of high-intensity focused ultrasound for treatment of periorbital, perioral, and neck wrinkles: Prospective open single-center single-arm confirmatory clinical trial.

Dermatologic therapy·2022
Same author

Induction of hair growth in hair follicle cells and organ cultures upon treatment with 30 kHz frequency inaudible sound via cell proliferation and antiapoptotic effects.

Biomedical reports·2022

相关实验视频

Updated: May 11, 2025

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

47.7K

在扩散MRI上基于深度学习的脑梗塞的自动细分.

Wi-Sun Ryu1,2, Dawid Schellingerhout3, Jonghyeok Park1

  • 1Artificial Intelligence Research Center, JLK Inc., Seoul, South Korea.

Scientific reports
|April 16, 2025
PubMed
概括

使用多站点数据和域调整训练深度学习模型显著改善了MRI上的脑梗塞细分. 使用大约2000张扩散权重图像 (DWI) 并进行适应,可以获得与更大的数据集可比的结果.

关键词:
深度学习是一种深度学习.扩散权重图像 扩散权重图像 扩散权重图像域名适应领域适应缺血性中风是因为缺血性中风.磁共振成像技术 磁共振成像技术

更多相关视频

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

8.8K
Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies
04:25

Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies

Published on: December 15, 2023

2.1K

相关实验视频

Last Updated: May 11, 2025

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

47.7K
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

8.8K
Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies
04:25

Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies

Published on: December 15, 2023

2.1K

科学领域:

  • 医疗成像医学成像
  • 人工智能的人工智能
  • 神经学 神经学

背景情况:

  • 在MRI上精确细分脑梗塞对于诊断和治疗至关重要.
  • 深度学习算法显示出希望,但需要大量,多样化的数据集以获得最佳性能.
  • 挑战包括不同医院站点和成像协议的数据变化.

研究的目的:

  • 评估培训数据大小和站点多样性对心脏病细分深度学习模型的影响.
  • 评估跨站点域调整技术的有效性.
  • 确定最佳策略,以提高算法通用性和性能.

主要方法:

  • 利用来自10所大学医院的10820张注释扩散权重图像 (DWI) 进行培训和内部测试.
  • 训练有素的基于3D U-net的算法,采用不同的样本大小 (217至8661个DWI).
  • 对独立数据集进行外部验证,并使用外部数据子集进行应用域调整.

主要成果:

  • 从217个多站点训练数据增加到1732个DWI,显著改善了子相似系数 (DSC) 和平均豪斯多夫距离 (AHD).
  • 数据大小的进一步增加产生了减少的回报,带来了边际的性能增长.
  • 仅使用50个外部图像进行域调整,在217个图像上训练的模型的表现与在8661个图像上训练的模型相提并论.

结论:

  • 多站点培训数据,大约2000个DWI,对于强大的心脏病细分至关重要.
  • 跨站点域名适应是提高性能和通用性的高效策略,特别是在有限的外部数据的情况下.
  • 深度学习模型从多样化,多站点数据和适应来获得可靠的脑梗塞细分的显著好处.