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

Prediction and intraoperative validation of pulmonary interlobar fissure development based on high-resolution thin-slice computed tomography and multiplanar reconstruction three-dimensional reconstruction technology: a prospective study.

Journal of thoracic disease·2026
Same author

Dual-wavelength laser-driven TCH degradation over ZnO/cu₁.₃₅O/g-C₃N₄ S-scheme heterojunctions: Charge-transfer dynamics, mineralization, and phytotoxicity assessment.

Journal of colloid and interface science·2026
Same author

Simultaneous inhibition of Microcystis growth, toxin production, and release by a high-efficacy algicidal bacterium Aeromonas sp. N3.

Journal of applied microbiology·2026
Same author

Genome-wide association studies and QTL mapping for traits deviating from normal distribution.

National science review·2026
Same author

Development and validation of a MODS risk prediction model for organophosphorus poisoning patients.

Frontiers in medicine·2026
Same author

Editorial: Highlights of 1st International Conference on Sustainable and Intelligent Phytoprotection (ICSIP 2025).

Frontiers in plant science·2026

相关实验视频

Updated: Jan 8, 2026

A Simple Protocol for Mapping the Plant Root System Architecture Traits
11:09

A Simple Protocol for Mapping the Plant Root System Architecture Traits

Published on: February 10, 2023

3.5K

KD-SSGD:知识蒸增强的半监督发芽检测检测.

Chengcheng Chen1, Di Luo1, Tiantian Pang2

  • 1School of Computer Science, Shenyang Aerospace University, Shenyang, China.

Frontiers in plant science
|December 24, 2025
PubMed
概括

这项研究引入了一种新的半监督框架,用于检测种子发芽,显著提高精度,最小的标记数据. 该方法为精准农业提供了高效的解决方案,减少了对大量数据注释的需求.

关键词:
深度学习是一种深度学习.组合学习组合学习发芽检测检测的发芽检测器知识的蒸知识的蒸.半监督物体检测半监督物体检测

更多相关视频

A Seed Coat Bedding Assay to Genetically Explore In Vitro How the Endosperm Controls Seed Germination in Arabidopsis thaliana
08:52

A Seed Coat Bedding Assay to Genetically Explore In Vitro How the Endosperm Controls Seed Germination in Arabidopsis thaliana

Published on: November 9, 2013

14.0K
Reliable Method for Assessing Seed Germination, Dormancy, and Mortality under Field Conditions
07:03

Reliable Method for Assessing Seed Germination, Dormancy, and Mortality under Field Conditions

Published on: November 6, 2016

10.9K

相关实验视频

Last Updated: Jan 8, 2026

A Simple Protocol for Mapping the Plant Root System Architecture Traits
11:09

A Simple Protocol for Mapping the Plant Root System Architecture Traits

Published on: February 10, 2023

3.5K
A Seed Coat Bedding Assay to Genetically Explore In Vitro How the Endosperm Controls Seed Germination in Arabidopsis thaliana
08:52

A Seed Coat Bedding Assay to Genetically Explore In Vitro How the Endosperm Controls Seed Germination in Arabidopsis thaliana

Published on: November 9, 2013

14.0K
Reliable Method for Assessing Seed Germination, Dormancy, and Mortality under Field Conditions
07:03

Reliable Method for Assessing Seed Germination, Dormancy, and Mortality under Field Conditions

Published on: November 6, 2016

10.9K

科学领域:

  • 农业科学 农业科学
  • 计算机视觉 计算机视觉
  • 机器学习 机器学习

背景情况:

  • 精准农业需要精确的种子发芽检测,用于作物监测和品种选择.
  • 完全监督的方法需要广泛的注释数据集,这在农业环境中是昂贵和耗时的.

研究的目的:

  • 为种子发芽检测开发一个高效的半监督学习框架,尽量减少对标记数据的依赖.
  • 引入一种新的知识蒸方法,使得无需预先培训的教师模型的端到端培训成为可能.

主要方法:

  • 一个包含轻量级蒸学生分支的教师-学生架构.
  • 关键模块包括用于伪标签优化的权重盒融合 (WBF),用于知识转移的特征蒸损失 (FDL) 和用于培训稳定性的分支适应权重 (BAW).

主要成果:

  • 在Maise-Germ数据集中仅使用1%的标记数据实现了47.0%的mAP,超过了现有的半监督方法.
  • 在三种谷物作物数据集上表现出强的表现,mAP在10%的标记数据上达到76.1%.
  • 在有限的监督下展示了强大的跨作物概括能力和有效的知识转移.

结论:

  • 该KD-SSGD框架提供高质量的伪标签和稳定,高精度检测与最小的标签数据.
  • 这种方法为智能农业感知和自动作物监测提供了高效和可扩展的解决方案.
  • 该方法显著降低了注释负担,使先进的计算机视觉技术更容易用于农业应用.