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

Add-On Chinese Medicine for Coronavirus Disease 2019 (ACCORD): A Retrospective Cohort Study of Hospital Registries.

The American journal of Chinese medicine·2021

Same author

Radiomics Features Predict <i>Telomerase Reverse Transcriptase</i> Promoter Mutations in World Health Organization Grade II Gliomas <i>via</i> a Machine-Learning Approach.

Frontiers in oncology·2021

Same author

Computation and data driven discovery of topological phononic materials.

Nature communications·2021

Same author

Deep learning applied to two-dimensional color Doppler flow imaging ultrasound images significantly improves diagnostic performance in the classification of breast masses: a multicenter study.

Chinese medical journal·2021

Same author

Shear viscosity prediction of alcohols, hydrocarbons, halogenated, carbonyl, nitrogen-containing, and sulfur compounds using the variable force fields.

The Journal of chemical physics·2021

Same author

Palladium-Catalyzed Atroposelective Coupling-Cyclization of 2-Isocyanobenzamides to Construct Axially Chiral 2-Aryl- and 2,3-Diarylquinazolinones.

Journal of the American Chemical Society·2021

Search research articles

相关实验视频

Updated: Jul 7, 2025

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

通过图像类别注释的弱监督突出物体检测.

Ruoqi Zhang¹, Xiaoming Huang¹, Qiang Zhu^1,2

¹Computer School, Beijing Information Science and Technology University, Beijing 100192, China.

Mathematical biosciences and engineering : MBE

|December 21, 2023

概括

本研究引入了一种弱监督的方法,用于仅使用类别标签检测突出物体. 它通过完善对象位置和生成高质量的伪标签来实现准确的检测,用于训练.

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:

深度学习已经推进了突出物体检测.
完全监督的方法需要广泛的像素级注释.
弱监督的方法使用更便宜的注释,如类别标签,但往往遭受不准确的检测.

研究的目的:

提出一种新的弱监督突出物体检测方法,仅使用类别注释.
为应对现有的基于类别的方法中不准确检测的挑战.
制定一种策略,用于生成和完善高质量的伪标签.

主要方法:

提出了一个粗略的对象定位网络 (COLN) 用于使用类别信息进行初始对象定位.
开发了一个伪标签生成和改进战略,包括基于标签对一致性的质量检查机制.
引入了一个多解码器神经网络 (MDN) 用于与生成的伪标签进行训练的突出检测.
实施了一种代伪标签更新策略,以优化伪标签和检测模型.

主要成果:

拟议的方法有效地从类别注释中生成像素级伪标签.
质量检查策略成功地选择了高度可靠的伪标签.
多解码器网络和代优化提高了检测准确度.

关键词:

深度学习是一种深度学习.图像类别注释图像类别注释标签检测标签检测标签检测标签检测突出的物体检测检测突出的物体检测监管能力较弱的监管机构.

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

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

Creating Objects and Object Categories for Studying Perception and Perceptual Learning

Creating Objects and Object Categories for Studying Perception and Perceptual Learning

Published on: November 2, 2012

相关实验视频

Last Updated: Jul 7, 2025

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

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

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

Creating Objects and Object Categories for Studying Perception and Perceptual Learning

Creating Objects and Object Categories for Studying Perception and Perceptual Learning

Published on: November 2, 2012

对四个公共数据集的评估表明,与现有的基于类别注释的方法相比,性能优越.

结论:

开发的弱监督方法为检测突出物体的完全监督方法提供了可行的替代方案.
类别注释,当与有效的伪标签策略相结合时,可以产生具有竞争力的检测性能.
拟议的方法显著推进了弱监督突出物体检测领域.