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

Independent and interactive effects of personal light exposure and air pollution on incident COPD: A prospective cohort study with 13 million hours of light sensor data.

Environmental pollution (Barking, Essex : 1987)·2026

Same author

An IGF2BP3-dependent metabolic circuit governs macrophage recruitment and immunosuppression in glioblastoma.

Cell reports·2026

Same author

Astaxanthin ameliorates cardiac damage induced by prepubertal di(2-ethylhexyl) phthalate exposure via inhibition of PPAR-α-mediated excessive autophagy.

Journal of environmental sciences (China)·2026

Same author

RBFOX2 suppresses NETosis and glioma growth via 5hmC-dependent PDGFB decay.

Cell reports·2026

Same author

An interpretable deep learning diagnostic framework for early Alzheimer's disease based on EEG microstate spectra and multi-branch CNN.

Journal of neuroengineering and rehabilitation·2026

Same author

Association between urban heat island and type 2 diabetes risk: genetic and proteomic insights.

Environment international·2026

Search research articles

相关实验视频

Updated: May 8, 2025

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

可变形的多层特征网络应用于核细分.

Shulei Chang¹, Tingting Yang¹, Bowen Yin¹

¹School of Information Science and Engineering, Shandong Normal University, Jinan, China.

Frontiers in microbiology

|December 24, 2024

概括

一个新的可变形多层特征网络 (DMFNet) 改善了医疗图像中的核细分. 这种方法提高了疾病评估的准确性,克服了现有的核细分技术的局限性.

科学领域:

医疗图像分析医学图像分析
计算病理学计算病理学
医疗保健中的人工智能

背景情况:

准确的核细分对于医学诊断和疾病评估至关重要.
现有的细分方法与核多样性和多样化的染色条件作斗争,限制了临床使用.

研究的目的:

引入一种新的可变形多层特征网络 (DMFNet),以改善核细分.
解决当前处理核变异性和染色差异方法的局限性.

主要方法:

拟议的DMFNet采用双层方法进行特征处理和面具生成.
可变形的卷积增强了特征提取,而平衡的特征金字塔则集成了多个尺度的特征.
一个单阶段实例细分框架直接根据位置生成面具.

主要成果:

在MoNuSeg 2018数据集上,DMFNet实现了37.8%的平均平均精度 (mAP) 和47.4%的平均平均回忆 (mAR).
性能超过了几种先进的核细分方法.
废弃性研究证实了个别网络模块的有效性.

结论:

DMFNet为医学成像中的核细分提供了强大而有效的解决方案.

关键词:

卷积神经网络是一种卷积神经网络.深度学习是一种深度学习.可变形多层特征网络的可变形多层特征网络核心细分的细分是核心的细分.病理学图像病理学图像

更多相关视频

Exploiting Live Imaging to Track Nuclei During Myoblast Differentiation and Fusion

Exploiting Live Imaging to Track Nuclei During Myoblast Differentiation and Fusion

Published on: April 13, 2019

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

相关实验视频

Last Updated: May 8, 2025

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

Exploiting Live Imaging to Track Nuclei During Myoblast Differentiation and Fusion

Exploiting Live Imaging to Track Nuclei During Myoblast Differentiation and Fusion

Published on: April 13, 2019

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

该网络显示出在医学图像分析和数字病理学方面的应用潜力很大.