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

A spectral-decomposition aging index for microplastics in aquatic environments.

Journal of hazardous materials·2026
Same author

Nitrifiers Drive Different N<sub>2</sub>O Production Patterns in Tropical River Sediments.

Environmental science & technology·2026
Same author

Associations between diet quality and multiple chemical exposures.

Ecotoxicology and environmental safety·2026
Same author

Shaping of microbial community structure and function by effluent-derived dissolved organic matter in a receiving river across seasons.

Eco-Environment & Health·2026
Same author

Microbial Perspective: Regulatory Mechanisms of Interactions Between Microplastics and Dissolved Organic Matter on Greenhouse Gas Emissions in Aquatic Ecosystems.

Global change biology·2026
Same author

Corrigendum to 'Exploring nonlinear responses of hydrodynamics and algal biomass to restoration measures in a eutrophic shallow urban lake' [Journal of Contaminant Hydrology 281 (2026), 104986].

Journal of contaminant hydrology·2026

相关实验视频

Updated: Jun 29, 2025

Dual-mode Imaging of Cutaneous Tissue Oxygenation and Vascular Function
11:35

Dual-mode Imaging of Cutaneous Tissue Oxygenation and Vascular Function

Published on: December 8, 2010

16.6K

双通道原型网络为少数镜头病理图像分类图像分类.

Hao Quan, Xinjia Li, Dayu Hu

    IEEE journal of biomedical and health informatics
    |April 8, 2024
    PubMed
    概括

    一个新的双通道原型网络 (DCPN) 有效地使用少量射击学习对病理图像进行分类,以有限的数据优于现有的罕见疾病诊断方法.

    科学领域:

    • 计算病理学计算病理学
    • 医学图像分析 医学图像分析
    • 人工智能在诊断中的应用

    背景情况:

    • 由于罕见疾病和注释挑战,有限的高质量数据集阻碍了病理学的深度学习.
    • 对数据稀缺的场景来说,近距离学习是有前途的,但在病理图像分类中未被充分探索.
    • 现有的方法在复杂的病理特征提取所需的概括性和精度方面扎.

    研究的目的:

    • 引入一种新的双通道原型网络 (DCPN),以高效地对病理图像进行少数镜头分类.
    • 通过提取多尺度,高精度的病理特征来提高原型表示的概括性.
    • 提高模型在复杂病理学图像分类任务中的辨别能力,特别是对于罕见疾病.

    主要方法:

    • 开发了一种双通道原型网络 (DCPN),将自主监督学习与金字塔视觉转换器 (PVT) 和卷积神经网络 (CNN) 集成在一起.
    • 利用基于多尺度特征的软投票分类器来提高模型性能.
    • 在三个公共数据集 (CRCTP,NCTCRC,LC25000) 上评估了DCPN,使用具有不同域移位的少数镜头分类任务.

    主要成果:

    • 在所有几次射击学习指标 (1次射击,5次射击,10次射击) 中,DCPN显著超过了原型网络.
    • 在同一个领域的任务中实现了最高的准确性:70.86% (1次拍摄),82.57% (5次拍摄) 和85.2% (10次拍摄),对原型网络的改进高达6.81%.

    更多相关视频

    Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging
    11:38

    Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging

    Published on: October 4, 2024

    548
    Multiplexed Barcoding Image Analysis for Immunoprofiling and Spatial Mapping Characterization in the Single-Cell Analysis of Paraffin Tissue Samples
    08:18

    Multiplexed Barcoding Image Analysis for Immunoprofiling and Spatial Mapping Characterization in the Single-Cell Analysis of Paraffin Tissue Samples

    Published on: April 7, 2023

    1.6K

    相关实验视频

    Last Updated: Jun 29, 2025

    Dual-mode Imaging of Cutaneous Tissue Oxygenation and Vascular Function
    11:35

    Dual-mode Imaging of Cutaneous Tissue Oxygenation and Vascular Function

    Published on: December 8, 2010

    16.6K
    Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging
    11:38

    Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging

    Published on: October 4, 2024

    548
    Multiplexed Barcoding Image Analysis for Immunoprofiling and Spatial Mapping Characterization in the Single-Cell Analysis of Paraffin Tissue Samples
    08:18

    Multiplexed Barcoding Image Analysis for Immunoprofiling and Spatial Mapping Characterization in the Single-Cell Analysis of Paraffin Tissue Samples

    Published on: April 7, 2023

    1.6K
  • 在10次射击设置中,DCPN的准确性 (85.2%) 超过了基于PVT的监督模型 (85.15%),证明了其用于罕见疾病诊断的潜力.

    • 结论:

    • DCPN是用于病理图像分类的高效的几次学习方法,在数据有限的情况下特别有价值.
    • 双通道架构和多尺度特征提取增强了模型的概括性和分类准确性.
    • 该DCPN显示显著的前景,以推进深度学习辅助的诊断在病理学,包括罕见疾病的识别.