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

Multi-strategy investigation identifies amentoflavone as a novel D3R antagonist for Parkinson's disease treatment.

Phytomedicine : international journal of phytotherapy and phytopharmacology·2026
Same author

ITM2A promotes thyroid cancer differentiation through metabolic reprogramming and enhances PD-L1-dependent T-cell responses.

Clinical and experimental medicine·2026
Same author

Duration-dependent effects of sucrose intake on fear extinction via distinct mechanisms in the amygdala and hippocampus.

Brain research·2026
Same author

Aptamer-functionalized Fe<sub>3</sub>O<sub>4</sub>-BR@Hemin nanozyme for the ultrasensitive colorimetric detection toward Staphylococcus aureus.

Talanta·2026
Same author

Associations of diabetes mellitus with/without diabetic retinopathy and cognitive outcomes in older adults: the potential role of the dietary inflammatory index in a multi-dataset observational study.

Frontiers in nutrition·2026
Same author

Targeting TLR4/non-canonical NF-κB pathway by Bushen Bitong recipe enhances BMSCs-based cartilage repair in osteoarthritis.

Biochemistry and biophysics reports·2026

相关实验视频

Updated: Jun 26, 2025

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

522

可扩展-RCNN:朝着高效率的增量少数射击物体检测方向.

Yiting Li1, Sichao Tian2, Haiyue Zhu3

  • 1Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore.

Frontiers in artificial intelligence
|May 8, 2024
PubMed
概括

本研究介绍了可扩展RCNN用于增量几次射击物体检测 (iFSOD). 它可以在不需要基础网络重新训练的情况下增加在线课程,超过现有方法的性能.

关键词:
几次射击的学习学习增量学习是一种增量学习.长尾的认可 长尾的认可对象检测检测对象检测对象检测零射击学习的学习

更多相关视频

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

1.8K
Long-term Video Tracking of Cohoused Aquatic Animals: A Case Study of the Daily Locomotor Activity of the Norway Lobster Nephrops norvegicus
05:57

Long-term Video Tracking of Cohoused Aquatic Animals: A Case Study of the Daily Locomotor Activity of the Norway Lobster Nephrops norvegicus

Published on: April 8, 2019

6.8K

相关实验视频

Last Updated: Jun 26, 2025

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

522
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

1.8K
Long-term Video Tracking of Cohoused Aquatic Animals: A Case Study of the Daily Locomotor Activity of the Norway Lobster Nephrops norvegicus
05:57

Long-term Video Tracking of Cohoused Aquatic Animals: A Case Study of the Daily Locomotor Activity of the Norway Lobster Nephrops norvegicus

Published on: April 8, 2019

6.8K

科学领域:

  • 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:

  • 增量短拍对象检测 (iFSOD) 在使用有限数据顺序学习新对象类别时提出了挑战.
  • 现有的方法在自适应检测系统中扎着过度装配和类别偏差.

研究的目的:

  • 提出一个高效和简单的框架,可扩展RCNN,以解决iFSOD问题.
  • 允许在线连续添加新类,而无需重新训练基础网络.

主要方法:

  • 调整了Faster R-CNN与两个新的组件:一个IOU-aware重量印记策略和一个小组软最大层 (GSL) 偏差校正.
  • 负债表意识的印记通过直接确定新类和背景的分类器重量来避免过度拟合.
  • 该GSL模块校准偏见的预测,以提高分类性能,防止灾难性遗忘.

主要成果:

  • 可扩展-RCNN显示了MS-COCO数据集的显著改进.
  • 拟议的方法在几次射击课程中表现比最先进的方法ONCE优于5.9分.
  • 实现了有效的在线自适应检测,基本网络的零重新训练.

结论:

  • 可扩展RCNN为iFSOD问题提供了有效的解决方案,使新对象类别的高效在线学习成为可能.
  • 负债表意识到重量印记和GSL的组合成功地减轻了过度匹配和类别偏差.
  • 该框架为需要顺序学习能力的自适应检测系统提供了强大的方法.