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

Concepts and Prototypes01:24

Concepts and Prototypes

327
The human nervous system handles vast amounts of information by translating sensory stimuli into neural impulses, which the brain processes, creating thoughts expressed through language or stored as memories. The brain also synthesizes information from emotions and memories, which significantly influence thoughts and behaviors. This intricate process creates a comprehensive mental picture.
The brain organizes this information using concepts, which are mental categories grouping linguistic data,...
327

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

Neural mechanisms of cough: insights from multimodal magnetic resonance imaging-a narrative review.

Journal of thoracic disease·2026
Same author

Ectopic lipid deposition in kidney diseases: mechanisms in specific cell types and therapeutic strategies.

Frontiers in endocrinology·2026
Same author

Anti-inflammatory and antioxidant effects of salidroside in diabetic nephropathy: a systematic review and meta-analysis of preclinical studies.

Frontiers in pharmacology·2026
Same author

Interferon-stimulated gene 15 small interfering RNA-loaded polarized mesoporous silica nanocarriers remodel the immune microenvironment to ameliorate post-myocardial infarction heart failure.

Materials today. Bio·2026
Same author

Association of the skeletal muscle mass to visceral fat area ratio with metabolically healthy obesity and metabolically unhealthy non-obesity: a cross-sectional study based on NHANES 2011-2018.

Lipids in health and disease·2025
Same author

<i>Lactiplantibacillus plantarum</i>-based inoculated fermentation of Guiyang sour rice-noodles improved its flavor quality.

Food chemistry: X·2025

相关实验视频

Updated: Nov 15, 2025

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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

369

不确定性引导的半监督的少数镜头细分与原型级核聚变.

Hailing Wang1, Chunwei Wu1, Hai Zhang1

  • 1Shanghai Key Laboratory of Trustworthy Computing, East China Normal University, Shanghai 200062, China; MOE Research Center for Software/Hardware Co-Design Engineering, East China Normal University, Shanghai 200062, China.

Neural networks : the official journal of the International Neural Network Society
|November 1, 2024
PubMed
概括

本研究介绍了以不确定性为指导的自适应原型网络 (UGAPNet),通过生成可靠的伪原型来改进少量射击的语义细分. 这种方法有效地减少了类内差异,并提高了新类别的细分精度.

关键词:
短暂的语义细分是短暂的语义细分原型学习学习的原型.原型级别的核聚变战略半监督学习 半监督学习不确定性 不确定性

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

2.7K
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

8.6K

相关实验视频

Last Updated: Nov 15, 2025

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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

369
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

2.7K
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

8.6K

科学领域:

  • 计算机视觉 计算机视觉
  • 机器学习 机器学习
  • 人工智能的人工智能

背景情况:

  • 短拍语义细分 (FSS) 由于有限的数据和高的类内差异,难以对新类别进行细分.
  • 将元知识转移到未见的类别是具有挑战性的,特别是在不同的支持-查询对中.

研究的目的:

  • 提出以不确定性为导向的自适应原型网络 (UGAPNet) 进行半监督的少数镜头语义细分.
  • 通过生成可靠的伪原型来解决类内部的语义偏见.
  • 通过处理可见和不可见的类来增强通用的少数镜头语义细分.

主要方法:

  • 使用共享的元学习器在未标记的图像上进行伪标签预测.
  • 包含一个不确定性估计模块,通过量化原型差异来消除伪标签.
  • 介绍了用于有效伪原型的原型纠正模块和通用自适应原型.
  • 建议一个原型级融合战略在原型的对比空间一般化FSS.

主要成果:

  • UGAPNet有效地生成可靠的伪原型,减轻类内语义偏差.
  • 不确定性估计和原型纠正模块提高了伪标签的质量.
  • 原型级融合战略成功地解决了一般化FSS中的混乱区域.
  • 对基准的实验证明了UGAPNet和融合战略的显著有效性.

结论:

  • UGAPNet提供了一种强大的解决方案,用于半监督的少数镜头语义细分,特别是在具有高类内差异的具有挑战性的场景中.
  • 拟议的方法增强了同时对可见和不可见类进行细分的能力.
  • 开发的技术在基准数据集上表现出强的表现,推动了FSS领域的发展.