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

Vision01:24

Vision

59.3K
Vision is the result of light being detected and transduced into neural signals by the retina of the eye. This information is then further analyzed and interpreted by the brain. First, light enters the front of the eye and is focused by the cornea and lens onto the retina—a thin sheet of neural tissue lining the back of the eye. Because of refraction through the convex lens of the eye, images are projected onto the retina upside-down and reversed.
59.3K

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

Renal Denervation in Cardiovascular Diseases: Mechanisms, Evidence, and Expanding Applications.

Catheterization and cardiovascular interventions : official journal of the Society for Cardiac Angiography & Interventions·2026
Same author

Intensive Versus Standard Systolic Blood Pressure Targets Following Successful Endovascular Thrombectomy for Acute Ischemic Stroke: A Systematic Review With Frequentist and Bayesian Meta-Analysis.

Stroke (Hoboken, N.J.)·2026
Same author

Comparative Outcomes of Transcatheter Edge-to-Edge Repair and Surgical Mitral Valve Repair or Replacement for Degenerative Mitral Regurgitation: A Systematic Review, Meta-Analysis, and Trial Sequential Analysis.

Catheterization and cardiovascular interventions : official journal of the Society for Cardiac Angiography & Interventions·2026
Same author

Efficacy and Safety of Aspirin-free versus Aspirin-based Strategies in Patients With Acute Coronary Syndrome Undergoing Percutaneous Coronary Intervention: A Systematic Review and Meta-Analysis.

Journal of cardiovascular pharmacology·2026
Same author

Awareness, Willingness, and Concerns about Clinical Trial Participation among Iraqi Patients: A Cross-Sectional Study.

Global heart·2026
Same author

Pulse Steroid Therapy in a Patient With Refractory Rhabdomyolysis of Unknown Etiology: A Case Report.

Clinical case reports·2026

相关实验视频

Updated: Jan 11, 2026

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

2.3K

眼科图像分类的高效视觉转换器:对监督,半监督和无监督学习方法的比较研究.

Ahmed Shakir Al-Wassiti1, Mohammed Tareq Mutar2, Ahmed Sermed Al Sakini3

  • 1MBChB, FIBMS (ophthalmology), FICO, FRCS (Glasg), College of Medicine, University of Baghdad, Baghdad, Baghdad Governorate, Iraq.

Journal of medical systems
|November 16, 2025
PubMed
概括

这项研究使用人工智能增强了眼科图像分类,结合了监督,半监督和无监督的学习,以最少的标记数据来改善诊断. MaxViT-L显示出有希望的性能,平衡精度和通用性,用于自动检测眼病.

关键词:
眼科成像 眼科成像眼科医生 眼科 眼科半监督学习 半监督学习没有监督的学习学习.视觉变压器 视觉变压器

更多相关视频

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
07:11

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping

Published on: December 8, 2023

2.3K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

729

相关实验视频

Last Updated: Jan 11, 2026

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

2.3K
Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
07:11

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping

Published on: December 8, 2023

2.3K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

729

科学领域:

  • 眼科医生 眼科 眼科
  • 计算机科学 计算机科学
  • 人工智能的人工智能
  • 机器学习 机器学习

背景情况:

  • 医学成像,特别是眼科,面临着高标注成本的挑战,限制了AI诊断工具的发展.
  • 监督学习模型需要大量的标记数据,而在眼科成像等专业领域,这些数据是稀缺而昂贵的.
  • 开发强大的诊断系统需要方法,即使使用有限的标记数据集,也可以有效执行.

研究的目的:

  • 调查监督,半监督和无监督学习策略的整合,以在标签稀缺条件下对眼科图像进行分类.
  • 通过利用最小的标记数据和强大的特征表示来提高眼科诊断性能.
  • 评估不同变压器架构和学习策略在改进人工智能驱动的眼科诊断方面的有效性.

主要方法:

  • 采用了18767张多式眼科图像的数据集 (1877张有标签,16890张没有标签).
  • 使用ViT-Base,DeiT-Base和MaxViT-L变压器架构进行员工监督学习.
  • 通过伪标签 (信心值≥0.98) 实现半监督学习,并使用基于SimCLR的对比学习和K-means集群进行无监督学习.

主要成果:

  • 在监督学习中,ViT-Base实现了92.47%的准确性. 在半监督伪标签后,MaxViT-L达到97.49%的准确性和0.9982 AUC.
  • 无监督对比学习与MaxViT-L改进的特征集群 (轮得分:0.556,DBI:0.541).
  • MaxViT-L在外部验证集上表现出卓越的性能,尽管计算复杂性更高,但在准确性和概括性之间提供了有利的权衡.

结论:

  • MaxViT-L,特别是在半监督和无监督的环境中,在诊断性能和眼科图像分类的模型概括之间提供了强大的平衡.
  • 综合方法有效地减少了对专家注释的依赖,为可扩展和自动化的眼科诊断解决方案铺平了道路.
  • 这项研究强调了先进的机器学习技术的潜力,以克服医疗AI中的数据稀缺挑战,提高眼科诊断能力.