Dataset of TEM Images for Carbon Nanomaterial Classification

We employed a convolutional neural network (CNN) for a machine learning and computer vision approach to recognize and classify airborne CNT (carbon nanotubes) and CNF (carbon nanofibers) particles from transmission electron microscope (TEM) images. The model we developed creates a framework for the automatic detection and classification of complex carbon nanostructures, with potential applications in the automated structural classification of other nanomaterials. This dataset comprises 5,323 grayscale TEM images of various nanostructured carbon samples classified into five categories: Cluster, Fiber, Matrix, Matrix-Surface, and Non-CNT. For more information about the data collection and classification process, please refer to https://doi.org/10.1039/D0NA00634C.本研究采用卷积神经网络（convolutional neural network, CNN）结合机器学习与计算机视觉技术，对透射电子显微镜（transmission electron microscope, TEM）图像中的空气中碳纳米管（carbon nanotubes, CNT）与碳纳米纤维（carbon nanofibers, CNF）颗粒进行识别与分类。本研究所构建的模型可为复杂碳纳米结构的自动检测与分类提供框架，有望拓展至其他纳米材料的自动化结构分类应用。本数据集包含5323张不同纳米结构碳样品的灰度TEM图像，共分为5个类别：团聚体（Cluster）、纤维（Fiber）、基体（Matrix）、基体-表面（Matrix-Surface）以及非碳纳米管（Non-CNT）。如需了解数据采集与分类流程的更多细节，请参考https://doi.org/10.1039/D0NA00634C。

搜集汇总

背景与挑战

背景概述

该数据集包含5,323张灰度TEM图像，用于碳纳米材料分类，分为Cluster、Fiber、Matrix、Matrix-Surface和Non-CNT五类。它通过卷积神经网络（CNN）实现机器学习和计算机视觉方法，旨在自动检测和分类复杂碳纳米结构，适用于纳米材料自动化结构分类应用。