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

Masking and Demasking Agents01:19

Masking and Demasking Agents

3.4K
EDTA titrations may necessitate masking and demasking agents to temporarily protect a particular metal ion in a mixture from the EDTA reaction. These agents facilitate the sequential analysis of the metal ions by forming stable complexes with some—but not all—metal ions during certain steps.
There are many masking agents, such as cyanide, fluoride, triethanolamine, thiourea, and 2,3-bis(sulfanyl)propan-1-ol (formerly 2,3-dimercapto-1-propanol), with the masking agent chosen based on...
3.4K

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

Identification of the gene cluster for the dithiolopyrrolone antibiotic holomycin in Streptomyces clavuligerus.

Proceedings of the National Academy of Sciences of the United States of America·2010
Same author

Safety evaluation of tea (Camellia sinensis (L.) O. Kuntze) flower extract: assessment of mutagenicity, and acute and subchronic toxicity in rats.

Journal of ethnopharmacology·2010
Same author

Influences of soil properties and leaching on nickel toxicity to barley root elongation.

Ecotoxicology and environmental safety·2010
Same author

Effects of CO2 insufflation on cerebrum during endoscopic thyroidectomy in a porcine model.

Surgical endoscopy·2010
Same author

Plants' use of different nitrogen forms in response to crude oil contamination.

Environmental pollution (Barking, Essex : 1987)·2010
Same author

Overexpression of p35 in Min6 pancreatic beta cells induces a stressed neuron-like apoptosis.

Journal of the neurological sciences·2010

相关实验视频

Updated: Jan 17, 2026

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

1.0K

自我监督的视频伪装物体检测和突出的物体检测.

Haoke Xiao1, Lv Tang2, Bo Li3

  • 1vivo Mobile Communication Co., Ltd, Shanghai, China; Institute of Artificial Intelligence, Xiamen University, Xiamen, 361100, China.

Neural networks : the official journal of the International Neural Network Society
|September 20, 2025
PubMed
概括

这项研究引入了一种新的自我监督的方法,用于视频伪装/物体检测 (VCOD/VSOD). 这种方法有效地解开了运动和上下文信息,超过了现有的无监督技术.

关键词:
深度学习是一种深度学习.自主监督学习学习视频伪装的物体检测检测器视频突出物体检测显着物体检测

更多相关视频

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

9.6K

相关实验视频

Last Updated: Jan 17, 2026

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

1.0K
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

9.6K

科学领域:

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

背景情况:

  • 像VCOD/VSOD这样的视频任务在多媒体中至关重要.
  • 现有的方法很难有效地利用运动和上下文信息.
  • 运动和上下文之间的偏差可以限制当前模型的性能.

研究的目的:

  • 为处理VCOD/VSOD中的动作和上下文信息提出一个新的脱而出的视角.
  • 开发一种自我监督的方法来降低注释成本.
  • 通过利用分离的信息来提高VCOD/VSOD任务的性能.

主要方法:

  • 一个新的模型使用ContextNet和MotionNet单独处理运动和上下文信息.
  • 一个自我监督的自适应路由机制将分配给ContextNet或MotionNet.
  • 交叉监督用于以自我监督的方式训练细分网络.

主要成果:

  • 提出的自我监督的脱模型实现了卓越的性能.
  • 在VCOD和VSOD数据集上始终优于最先进的无监督方法.
  • 证明了解开运动和上下文信息的有效性.

结论:

  • 解的视角为VCOD/VSOD任务提供了一个有希望的方向.
  • 自主监督学习显著减少了注释要求.
  • 拟议的方法为无监督的VCOD/VSOD设定了一个新的基准.