Search research articles

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 Journal Methods Collections JoVE Encyclopedia of Experiments 存档

教育

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual 教师资源中心教师网站

使用条款与条件

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

Programmable Electrohydrodynamic Printing of Pt-CNTs Nanointerfaces via In Situ Thermal Regulation for High-Fidelity Biosensing.

ACS applied materials & interfaces·2026

Same author

Mitochondrial OXPHOS restricts SARS-CoV-2 replication.

Science advances·2026

Same author

The roles of gut microbiota and their metabolites in uric acid-related metabolic diseases: mechanisms and therapeutic targets.

Frontiers in microbiology·2026

Same author

Mitochondria and Qi: Merging eastern and western medicine.

Pharmacological research·2026

Same author

The dual functions of Ziziphi spinosae semen as medicine and food: origins, cross-domain applications, antidepressant metabolites, and pharmacological mechanisms.

Frontiers in pharmacology·2026

Same author

Spatiotemporally controlled systemic delivery reshapes the in vivo fate of a cationic antimicrobial peptide for lung-selective exposure and an improved therapeutic index.

European journal of pharmaceutics and biopharmaceutics : official journal of Arbeitsgemeinschaft fur Pharmazeutische Verfahrenstechnik e.V·2026

Same journal

Real-time EEG-based epileptic seizure prediction using artificial intelligence: A systematic review.

Artificial intelligence in medicine·2026

Same journal

R-peak detection and ECG data compression scheme based on empirical mode decomposition and wavelet transform.

Artificial intelligence in medicine·2026

Same journal

CastNet: A three-channel EEG-based deep learning model for cross-subject depression detection.

Artificial intelligence in medicine·2026

Same journal

State-of-the-art TinyML approaches for colorectal cancer detection: Current advances, challenges, and future directions.

Artificial intelligence in medicine·2026

Same journal

JRadiEvo: A Japanese radiology report generation model enhanced by evolutionary optimization of model merging.

Artificial intelligence in medicine·2026

Same journal

Causally-informed deep learning towards explainable and generalizable outcome prediction in critical care.

Artificial intelligence in medicine·2026

查看所有相关文章

Search research articles

相关实验视频

Updated: Jul 9, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

无源域自适应细分与类平衡的补充自我训练.

Yongsong Huang¹, Wanqing Xie², Mingzhen Li³

¹Harvard Medical School, Harvard University, Boston, MA, USA; Department of Communications Engineering, Graduate School of Engineering, Tohoku University, Sendai, Miyagi, Japan; Gordon Center for Medical Imaging, Massachusetts General Hospital, Boston, MA, USA.

Artificial intelligence in medicine

|December 2, 2023

概括

此摘要是机器生成的。

本研究引入了一个新的类平衡的补充自我训练 (CBCOST) 框架,用于源代码免费的无监督域调整 (SFUDA) 分段. 在没有源数据的情况下,CBCOST有效地解决了类不平衡和伪标签噪声,提高了细分精度.

关键词:

分段化分段化分段化分段化进行自我训练.没有源代码的域名调整.

更多相关视频

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

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

Published on: November 30, 2022

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

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

Published on: December 15, 2023

相关实验视频

Last Updated: Jul 9, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

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

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

Published on: November 30, 2022

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

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

Published on: December 15, 2023

科学领域:

计算机视觉和机器学习
医学图像分析医学图像分析

背景情况:

无监督域调整 (UDA) 对于将模型应用于新数据集至关重要,但由于隐私,标记源数据访问往往受到限制.
无源UDA (SFUDA) 提供了一个解决方案,但与阶级不平衡 ("赢家夺取一切") 和自我培训中的杂伪标签作斗争.
现有的SFUDA方法往往无法充分细分少数群体,并且容易产生过度自信的伪标签噪音.

研究的目的:

提出一个新的框架,类平衡的补充自我训练 (CBCOST),以克服无源无监督域调整 (SFUDA) 细分的局限性.
通过解决SFUDA中的类不平衡和伪标签噪声来提高细分性能.
为了使有效的知识从标记的源域转移到未标记的目标域,而无需在调整过程中使用源数据.

主要方法:

开发了一个CBCOST框架,通过两个关键组成部分共同优化伪标签自我训练:类智能平衡伪标签训练 (CBT) 和补充自我训练 (COST).
CBT利用细粒度的类智能信心和适应值来选择平衡的伪标签像素,减轻'赢家夺取全部'问题.
COST采用启发式补充标签选择来过不正确的伪标签,减少噪音并提高模型的稳定性.

主要成果:

与现有的SFUDA方法相比,CBCOST框架在各种细分任务中表现出更高的性能.
在2D/3D交叉模式心脏和脑瘤细分上进行评估,CBCOST取得了具有竞争力的结果,与使用源数据的传统UDA方法相美.
实验结果验证了CBCOST在处理类不平衡和伪标签噪声方面的有效性,以改善SFUDA细分.

结论:

拟议的CBCOST框架有效地解决了SFUDA中的关键挑战,即阶级不平衡和伪标签噪音.
当源数据不可用时,CBCOST为域调整提供了一个强大的解决方案,实现高分段精度.
这种方法显示了在医疗图像细分领域的真实应用的巨大潜力,数据隐私是一个问题.