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

ITGA5 is overexpressed and promotes tumor progression through SNAI2 in OSCC.

Frontiers in cell and developmental biology·2026

Same author

HFB301001, an OX40-based immunotherapy, drives Treg clearance and CTL activation through optimized OX40 receptor clustering.

Journal for immunotherapy of cancer·2026

Same author

Immunogenic tumor cell death and T-cell-derived IFN-γ elicit tumoricidal macrophages to potentiate OX40 immunotherapy.

Cell reports. Medicine·2026

Same author

Structural optimization and biological evaluation of 5-fluoro-7H-pyrrolo[2,3- d]pyrimidine derivatives as potent FAK inhibitors for the treatment of triple negative breast cancers (TNBC).

European journal of medicinal chemistry·2025

Same author

Evaluation of anti-liver cancer activity and anticancer mechanism of one novel small molecule compound (THY-10A62) targeting FAK pathway.

Frontiers in oncology·2025

Same author

Weakly supervised deep learning for multimodal MRI-TRUS registration: Toward assisting prostate biopsy guidance.

Digital health·2025

Search research articles

相关实验视频

Updated: May 21, 2025

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

隐私保护数据增强用于数字病理使用改进的DCGAN.

Fengjun Hu, Fan Wu, Dongping Zhang

IEEE journal of biomedical and health informatics

|March 18, 2025

概括

此摘要是机器生成的。

本研究介绍了一种改进的深度卷积生成对抗网络 (DCGAN),用于整张幻灯片图像 (WSI) 数据增强. 该方法通过生成高质量的合成图像来提高数字病理学的深度学习模型性能,这对于精准医学至关重要.

更多相关视频

Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures

Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures

Published on: August 5, 2021

Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment

Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment

Published on: June 3, 2022

相关实验视频

Last Updated: May 21, 2025

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

Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures

Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures

Published on: August 5, 2021

Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment

Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment

Published on: June 3, 2022

科学领域:

数字病理学数字病理学
人工智能在医学中的应用
计算病理学计算病理学

背景情况:

整片图像 (WSI) 分析对于精准医学至关重要,特别是在瘤学中.
由于隐私问题,WSI数据集的有限可用性阻碍了深度学习模型的开发和通用性.
当前的数据增强技术在生成现实和多样化的病理图像方面往往不足.

研究的目的:

建议使用深度卷积生成对抗网络 (DCGAN) 改进全幻灯片图像 (WSI) 的数据增强方法.
提高合成WSI数据的质量和多样性,以改善深度学习模型培训.
解决隐私法规对WSI数据集可用性的限制.

主要方法:

利用CTransPath进行自我监督预训,提取丰富的WSI特征,用于指导合成图像生成.
实施了改进的DCGAN,包括最小平方对抗损失和频域损失,以提高像素精度和结构保真度.
集成剩余块和跳过连接以加深网络,改善梯度流,稳定训练.

主要成果:

与PatchCamelyon数据集上的传统模型相比,增强的DCGAN获得了优异的结构相似度指数 (SSIM) 和Fréchet初始距离 (FID) 评分.
由拟议方法生成的增强数据集显著改善了下游分类任务的性能.
关键性能指标,包括准确性,曲线下的面积 (AUC) 和F1得分得到了大幅提高.

结论:

提议的改进的DCGAN有效生成高保真度的合成WSI数据,克服数据稀缺的局限性.
这种数据增强策略显著提高了数字病理学深度学习模型的性能和通用性.
该方法有望通过更强大的AI驱动的WSI分析来推进精密瘤学.