有效的半监督语义细分电子显微镜癌症图像的稀疏注释
Lucas Pagano1,2, Guillaume Thibault1, Walid Bousselham1
1Department of Biomedical Engineering, Oregon Health and Science University, Portland, OR, United States.
Frontiers in bioinformatics
|January 1, 2024
概括
深度学习模型通过自动化细胞核和细胞核的细分来显著加速用于癌症研究的电子显微镜 (EM) 图像的分析. 本研究比较了几种模型,突出了先进架构和半监督学习的好处.
科学领域:
- 生物医学成像技术 生物医学成像技术
- 计算病理学计算病理学
- 癌症研究 癌症研究
背景情况:
- 电子显微镜 (EM) 提供纳米分辨率成像,对于了解癌症治疗耐药性至关重要.
- 手动分析EM数据用于结构识别是一个重要的瓶,需要每个样本数月的时间.
- 深度学习 (DL) 为自动化和加快EM图像分析提供了一个有前途的解决方案.
研究的目的:
- 评估和比较最先进的深度学习模型,用于在3D瘤活检卷中细分细胞核和细胞核.
- 使用未标记的数据评估半监督学习 (交叉伪监督) 的有效性.
- 确定最有效的DL方法来缓解EM数据分析中手动细分瓶.
主要方法:
- 使用ResUNet,UNet++,FracTALResNet,SenFormer和CEECNet模型进行细分的比较.
- 实施半监督学习 (交叉伪监督) 使用未标记的数据.
- 在三个内部注释数据集中对稀疏的手册标签进行培训和评估.
主要成果:
- 在评估的深度学习模型中,证明了3D Dice得分的改善.
- 量化了更复杂的模型架构的性能增长.
- 展示了半监督学习在提高细分精度方面的附加值.
结论:
- 先进的深度学习模型比以前的EM图像细分方法提供了实质性的改进.
- 半监督学习有效地利用未标记的数据来提高细分性能.
- 这些发现为在癌症研究中更快,更有效地分析EM数据铺平了道路.
更多相关视频
13:01Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment
Published on: June 3, 2022
3.8K
04:48Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.8K
相关概念视频
01:20Immunogold Electron Microscopy
4.0K
Immunoelectron microscopy utilizes immunogold labeling of endogenous proteins with specific antibodies to detect and localize these proteins in cells and tissues. The procedure provides insights into the distribution and quantification of protein under different stimulation conditions offering clues about their functions. Conjugating highly electron-dense gold particles with primary or secondary antibodies allow antigen detection on and within cells, with high resolution and specificity.
4.0K
01:07Scanning Electron Microscopy
4.2K
A scanning electron microscope (SEM) is used to study the surface features of a sample by using an electron beam that scans the sample surface in a two-dimensional manner. Typically, areas between ~1 centimeter to 5 micrometers in width can be imaged. SEM can be used to image bacteria, viruses, tissues as well as larger samples like insects. Conventional SEM gives a magnification ranging from 20X to 30,000X and spatial resolution of 50 to 100 nanometers.
Fundamental Principles
Accelerated...
Fundamental Principles
Accelerated...
4.2K