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

Characterization of PANoptosis-related expression pattern, prognosis and tumor microenvironment in head and neck squamous cell carcinoma.

Discover oncology·2025
Same author

Analysis of the correlation between glycemic variability indices and disease severity as well as inflammatory status in patients with type 2 diabetes mellitus and periodontitis.

Clinical oral investigations·2025
Same author

Semi-Supervised Medical Image Segmentation Based on Frequency Domain Aware Stable Consistency Regularization.

Journal of imaging informatics in medicine·2025
Same author

Prediction of Freezing of Gait in Parkinson's disease based on multi-channel time-series neural network.

Artificial intelligence in medicine·2024
Same author

A pyroptosis-related gene expression signature predicts immune microenvironment and prognosis in head and neck squamous cell carcinoma.

European archives of oto-rhino-laryngology : official journal of the European Federation of Oto-Rhino-Laryngological Societies (EUFOS) : affiliated with the German Society for Oto-Rhino-Laryngology - Head and Neck Surgery·2023
Same author

A novel and independent survival prognostic model for OSCC: the functions and prognostic values of RNA-binding proteins.

European archives of oto-rhino-laryngology : official journal of the European Federation of Oto-Rhino-Laryngological Societies (EUFOS) : affiliated with the German Society for Oto-Rhino-Laryngology - Head and Neck Surgery·2023

相关实验视频

Updated: Jul 5, 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

556

一个快速的注意力引导的等级解码网络,用于实时语义细分.

Xuegang Hu1, Jing Feng2

  • 1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China.

Sensors (Basel, Switzerland)
|January 11, 2024
PubMed
概括
此摘要是机器生成的。

本研究介绍了FAHDNet,一个快速的注意力引导的等级解码网络,用于实时语义细分. 它实现了高精度和速度,改善了现场理解的实际应用.

关键词:
注意力机制注意力机制编码器解码器网络功能融合功能融合功能实时语义细分实时语义细分

更多相关视频

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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

405
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

2.8K

相关实验视频

Last Updated: Jul 5, 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

556
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

405
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

2.8K

科学领域:

  • 计算机视觉 计算机视觉
  • 人工智能的人工智能
  • 深度学习 (Deep Learning) 是一种深度学习.

背景情况:

  • 语义细分对于场景理解至关重要,但由于复杂的模型,它经常遭受缓慢的推理速度.
  • 实时应用需要高效的语义细分模型,平衡准确性和速度.

研究的目的:

  • 提出一个新的网络,FAHDNet,用于快速准确的实时语义细分.
  • 在现有的语义细分模型中解决准确性和推断速度之间的权衡问题.

主要方法:

  • 开发了一种不对称的U形网络 (FAHDNet),采用一个编码器和一个多尺度瓶残留单元 (MBRU),结合了注意力和分解卷积.
  • 引入了一个空间信息补偿 (SIC) 模块来恢复丢失的空间纹理信息.
  • 在解码器中实现了全局注意 (GA) 模块,以增强功能交互,以及用于多尺度功能集成的轻量级分级解码器.

主要成果:

  • 在Cityscapes数据集上,FAHDNet在135FPS实现了70.6%的mIoU.
  • 在Camvid数据集上,该网络在335 FPS时达到67.2%的mIoU.
  • 与现有网络相比,在准确性和推断速度方面都表现出卓越的性能.

结论:

  • FAHDNet有效地平衡了高精度与快速推断速度,用于实时语义细分.
  • 拟议的网络在要求快速场景理解的应用中增强了实际的可用性.
  • 注意力机制和层次解码的整合有助于改进多尺度对象细分.