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

IFNγ-Producing B Cells Play a Regulating Role in Infection-Mediated Inhibition of Allergy.

Biology·2023

Same author

Synthesis of hierarchically porous zirconium-based metal-organic framework@silica core-shell stationary phase through etching strategy for liquid chromatography.

Journal of chromatography. A·2023

Same author

Lumican silencing ameliorates β-glycerophosphate-mediated vascular smooth muscle cell calcification by attenuating the inhibition of APOB on KIF2C activity.

Open medicine (Warsaw, Poland)·2023

Same author

N6-methyladenosine-induced METTL1 promotes tumor proliferation via CDK4.

Biological chemistry·2023

Same author

Circ_0005736 promotes tenogenic differentiation of tendon-derived stem cells through the miR-636/MAPK1 axis.

Journal of orthopaedic surgery and research·2023

Same author

Genetic alterations in juvenile cervical clear cell adenocarcinoma unrelated to human papillomavirus.

Frontiers in medicine·2023

Search research articles

相关实验视频

Updated: Jul 7, 2025

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping

Published on: April 21, 2023

基于深度学习的6-DoF对象构成估计,考虑合成数据集.

Tianyu Zheng¹, Chunyan Zhang¹, Shengwen Zhang¹

¹School of Mechanical Engineer, Jiangsu University of Science and Technology, Zhenjiang 212100, China.

Sensors (Basel, Switzerland)

|December 23, 2023

概括

这项研究引入了一种新的深度学习方法,用于精确的6度自由度 (6-DoF) 对象姿势估计. 它弥合了合成与现实领域的差距,提高了机器人和计算机视觉应用的准确性和概括性.

科学领域:

计算机视觉计算机视觉
深度学习 (Deep Learning) 是一种深度学习.
机器人技术机器人技术机器人技术

背景情况:

生成高质量的6度自由度 (6-DoF) 对象姿势估计数据集是具有挑战性的.
合成和真实数据之间的域间隙限制了现有的姿势估计方法的准确性和概括性.

研究的目的:

提出一种新的方法来增强6-DoF对象构成估计准确性和概括性.
通过使用改进的数据集和深度学习技术来解决领域差距问题.

主要方法:

使用Blenderproc来生成高质量的合成数据,并使用双边过来处理以尽量减少域间隙.
开发了一种基于注意力的面具基于区域的卷积神经网络 (R-CNN),以提高检测精度和降低计算成本.
引入了改进的特征金字塔网络 (iFPN),增加了自下而上的路径,以增强特征提取.
提出了一种新型的卷积块注意力模块-卷积表示自编码器 (CBAM-CDAE) 网络,包含通道和空间注意力机制.
实现了姿势改进,以准确地确定6-DoF对象的姿势.

主要成果:

提出的基于注意力的面具R-CNN显著提高了检测准确度.
在iFPN有效地提取更深的图像特征.

关键词:

6-DoF对象姿势估计估计在CBAMCDAE上通过双边过进行过.深度学习是一种深度学习.合成数据集是一种合成数据集.

更多相关视频

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

Author Spotlight: Automated Deep Brain Stimulation for Parkinson's Disease - Exploring the Possibilities and Challenges of Home Monitoring

Author Spotlight: Automated Deep Brain Stimulation for Parkinson's Disease - Exploring the Possibilities and Challenges of Home Monitoring

Published on: July 14, 2023

相关实验视频

Last Updated: Jul 7, 2025

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping

Published on: April 21, 2023

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

Author Spotlight: Automated Deep Brain Stimulation for Parkinson's Disease - Exploring the Possibilities and Challenges of Home Monitoring

Author Spotlight: Automated Deep Brain Stimulation for Parkinson's Disease - Exploring the Possibilities and Challenges of Home Monitoring

Published on: July 14, 2023

该CBAM-CDAE增强了自动编码器的特征提取能力.

对T-LESS和LineMOD数据集的评估表明,与现有模型相比,它们的性能优越.

结论:

提出的方法有效地减少了合成数据和真实数据之间的领域差距.
新型的深度学习架构在6-DoF对象姿势估计中实现了最先进的准确性.
这种方法为需要精确的物体姿势信息的现实应用提供了有前途的解决方案.