Transmission electron microscopy (TEM) image datasets of peptide / protein nanowire morphologies

TEM image dataset containing four nanowire morphologies of bio-derived protein nanowires and synthetic peptide nanowires. The peptide / protein nanowires used in this study were synthesized and imaged by Brian Montz in Prof. Todd Emrick's research group at the Department of Polymer Science and Engineering Department, University of Massachusetts Amherst. We acknowledge financial support from the U.S. National Science Foundation, Grant NSF DMREF #1921839 and DMREF #1921871. Nanowires were classified into either of the four morphologies: bundle, singular, dispersed or network. Each morphology contains 100 images (jpg files). For the dispersed and network morphologies, because these two morphologies are harder to visually distinguish, we have created manual segmentation labels of the nanowires (included in these two morphology folders as png files). Percolation analysis was done on these manually segmented nanowires to provide quantitative metric on whether the nanowires form a network in the image. seg_mask_5_resolutions.zip contains ground truth 2D binary encoding of segmented nanowires at 5 resolutions. encoders_trained_with_optimized_hyperparameter.zip contains 4 sets of encoders trained with either SimCLR or Barlow-Twins self-supervised methods on either generic TEM images, or generic everyday photographic images (each with 5 replicates with different random seed) with optimized hyperparameters. Open-access datasets that have been used during self-supervised training. 2021-CEM500K.zip contains 10,000 images that was used as "generic TEM images" to train the encoders with self-supervised methods, these are a random selection from the CEM500k open-access dataset. DOI: 10.7554/eLife.65894 2022-1000-ImageNet.zip contains 1,000 images from the ImageNet1k dataset, each come from a different category. DOI: 10.1007/s11263-015-0816-y Open-access datasets that our machine learning workflow have been applied to: 2022-AutoDetect-mNP-morphology.zip contains a selected TEM images of nanoparticles categorized in 3 morphologies from the AutoDetect-mNP datasets: DOI: 10.6078/D1WT44 and DOI: 10.6078/D1S12H 2021-TEM virus.zip contains TEM images of 9 types of viruses from the TEM virus dataset. Matuszewski, Damian; Sintorn, Ida-Maria (2021), “TEM virus dataset”, Mendeley Data, V3, DOI: 10.17632/x4dwwfwtw3.3 The official github page of the implementation of the machine learning models is semi-supervised_learning_microscopy_images. If you use the dataset or the codes in the repository linked above, please cite the following manuscript: S. Lu, B. Montz, T. Emrick and A. Jayaraman, Digital Discovery, 2022, 1, 816-833 , DOI: 10.1039/D2DD00066K本数据集为透射电子显微镜（Transmission Electron Microscopy, TEM）数据集，包含生物源蛋白质纳米线与合成肽纳米线共4种纳米线形貌。 本研究使用的肽/蛋白质纳米线由马萨诸塞大学阿默斯特分校聚合物科学与工程系Todd Emrick教授研究组的Brian Montz合成并完成成像。 本研究感谢美国国家科学基金会（National Science Foundation, NSF）的资助，资助编号为NSF DMREF #1921839与DMREF #1921871。 纳米线被划分为以下4种形貌：束状、单根、分散态与网状。每种形貌包含100张jpg格式图像。 由于分散态与网状形貌较难通过视觉区分，我们为这两类形貌制作了纳米线的人工分割标签（以png文件形式存储于对应形貌文件夹中）。我们对这些人工分割后的纳米线进行了渗流分析，以量化指标评估图像中纳米线是否形成网状结构。 seg_mask_5_resolutions.zip 包含5种分辨率下的纳米线分割结果的二维二进制编码真值标签。 encoders_trained_with_optimized_hyperparameter.zip 包含4组经优化超参数训练得到的编码器，这些编码器分别基于SimCLR或Barlow-Twins自监督方法，训练数据集既可为通用透射电子显微镜图像，也可为通用日常摄影图像（每组均使用5种不同随机种子进行重复实验）。 自监督训练过程中使用的公开数据集： 2021-CEM500K.zip 包含10000张图像，用作自监督方法训练编码器的“通用透射电子显微镜图像”，这些图像选自CEM500k公开数据集的随机子集，其DOI为：10.7554/eLife.65894。 2022-1000-ImageNet.zip 包含来自ImageNet1k数据集的1000张图像，每张图像对应一个不同的类别，其DOI为：10.1007/s11263-015-0816-y。 本机器学习工作流程所应用的公开数据集： 2022-AutoDetect-mNP-morphology.zip 包含选自AutoDetect-mNP数据集的、按3种形貌分类的纳米颗粒透射电子显微镜图像，对应DOI为：10.6078/D1WT44 与 10.6078/D1S12H。 2021-TEM virus.zip 包含来自TEM病毒数据集的9种病毒的透射电子显微镜图像。相关数据集信息：Matuszewski, Damian; Sintorn, Ida-Maria (2021), “TEM virus dataset”, Mendeley Data, V3, DOI: 10.17632/x4dwwfwtw3.3。 本机器学习模型实现的官方GitHub页面为 semi-supervised_learning_microscopy_images。 若您使用本数据集或上述链接的代码仓库，请引用以下论文：S. Lu, B. Montz, T. Emrick and A. Jayaraman, Digital Discovery, 2022, 1, 816-833, DOI: 10.1039/D2DD00066K。