Trainable segmentation for transmission electron microscope images of inorganic nanoparticles

Abstract

We present a trainable segmentation method implemented within the python package ParticleSpy. The method takes user labelled pixels, which are used to train a classifier and segment images of inorganic nanoparticles from transmission electron microscope images. This implementation is based on the trainable Waikato Environment for Knowledge Analysis (WEKA) segmentation, but is written in python, allowing a large degree of flexibility and meaning it can be easily expanded using other python packages. We find that trainable segmentation offers better accuracy than global or local thresholding methods and requires as few as 100 user-labelled pixels to produce an accurate segmentation. Trainable segmentation presents a balance of accuracy and training time between global/local thresholding and neural networks, when used on transmission electron microscope images of nanoparticles. We also quantitatively investigate the effectiveness of the components of trainable segmentation, its filter kernels and classifiers, in order to demonstrate the use cases for the different filter kernels in ParticleSpy and the most accurate classifiers for different data types. A set of filter kernels is identified that are effective in distinguishing particles from background but that retain dissimilar features. In terms of classifiers, we find that different classifiers perform optimally for different image contrast; specifically, a random forest classifier performs best for high-contrast ADF images, but that QDA and Gaussian Naïve Bayes classifiers perform better for low-contrast TEM images.

nNPipe: a neural network pipeline for automated analysis of morphologically diverse catalyst systems

“…However, steps including model generation, image simulation and augmentation, as well as training of the individual networks, require considerable processing time before experimental data can be analysed. For cases where only small datasets are considered, it might therefore be advantageous to use shallow network segmentation or other machine learning techniques which have been shown to achieve high performance for uniform sample morphologies such as the Au/Ge system considered here without a significant training overhead 37,50,51 .…”

Unsupervised Analysis of Optical Imaging Data for the Discovery of Reactivity Patterns in Metal Alloy

“…One common approach is the use of machine vision procedures that employ computer algorithms. [4,[30][31][32][33] While these methods are automated, they can be subject to human bias and may require finetuning of empirical parameters. In this study, we propose an efficient unsupervised machine learning framework for fully automated analysis of reflective videos obtained from the operando monitoring of the local electrochemical degradation by aqueous solution of Al alloy of 6061 series as a model system.…”

A Lens on the Future: Next Generation Microscopy by Next Generation Microscopists

“…We remain in the realm of electron microscopy with Bell et al, this time focusing on a trainable segmentation tool for quantifying the size/shape distribution of nanoparticles; ParticleSpy. 1 Bell and colleagues present a well-characterised, openly accessible, and flexible method for quantifying the dimensions of nanoparticles imaged by TEM and other modalities. The trainable and accountable nature of their analysis method allows the user to quickly perform training on a subset of data before allowing ParticleSpy to do its work.…”

“…Indeed, the articles gathered within this Issue demonstrate developments across the full breadth of the RMS community, from developments of open‐source codes enabling improved automated particle counting in (S)TEM 1 environmental AFM applied to dental enamel (2) and an introduction to THz methods as useful for biological materials 3 showing the wide range of our community and the interrelation of all of these techniques to each other.…”

“…We remain in the realm of electron microscopy with Bell et al. , this time focusing on a trainable segmentation tool for quantifying the size/shape distribution of nanoparticles; ParticleSpy 1 . Bell and colleagues present a well‐characterised, openly accessible, and flexible method for quantifying the dimensions of nanoparticles imaged by TEM and other modalities.…”