Calculating small-angle scattering intensity functions from electron-microscopy images.

We outline procedures to calculate small-angle scattering (SAS) intensity functions from 2-dimensional electron-microscopy (EM) images. Two types of scattering systems were considered: (a) the sample is a set of particles confined to a plane; or (b) the sample is modelled as parallel, infinitely long cylinders that extend into the image plane. In each case, an EM image is segmented into particle instances and the background, whereby coordinates and morphological parameters are computed and used to calculate the constituents of the SAS-intensity function. We compare our results with experimental SAS data, discuss limitations, both general and case specific, and outline some applications of this method which could potentially complement experimental SAS.

Bayesian Particle Instance Segmentation for Electron Microscopy Image Quantification.

Automating the analysis portion of materials characterization by electron microscopy (EM) has the potential to accelerate the process of scientific discovery. To this end, we present a Bayesian deep-learning model for semantic segmentation and localization of particle instances in EM images. These segmentations can subsequently be used to compute quantitative measures such as particle-size distributions, radial- distribution functions, average sizes, and aspect ratios of the particles in an image. Moreover, by making use of the epistemic uncertainty of our model, we obtain uncertainty estimates of its outputs and use these to filter out false-positive predictions and hence produce more accurate quantitative measures. We incorporate our method into the ImageDataExtractor package, as ImageDataExtractor 2.0, which affords a full pipeline to automatically extract particle information for large-scale data-driven materials discovery. Finally, we present and make publicly available the Electron Microscopy Particle Segmentation (EMPS) data set. This is the first human-labeled particle instance segmentation data set, consisting of 465 EM images and their corresponding semantic instance segmentation maps.

3-D Inorganic Crystal Structure Generation and Property Prediction via Representation Learning.

Generative models have been successfully used to synthesize completely novel images, text, music, and speech. As such, they present an exciting opportunity for the design of new materials for functional applications. So far, generative deep-learning methods applied to molecular and drug discovery have yet to produce stable and novel 3-D crystal structures across multiple material classes. To that end, we, herein, present an autoencoder-based generative deep-representation learning pipeline for geometrically optimized 3-D crystal structures that simultaneously predicts the values of eight target properties. The system is highly general, as demonstrated through creation of novel materials from three separate material classes: binary alloys, ternary perovskites, and Heusler compounds. Comparison of these generated structures to those optimized via electronic-structure calculations shows that our generated materials are valid and geometrically optimized.

ImageDataExtractor: A Tool To Extract and Quantify Data from Microscopy Images.

The rise of data science is leading to new paradigms in data-driven materials discovery. This carries an essential notion that large data sources containing chemical structure and property information can be mined in a fashion that detects and exploits structure-property relationships, such that chemicals can be predicted to suit a given material application. The success of material predictions is predicated on these large data sources of chemical structure and property information being suited to a target application. Microscopy is commonly used to characterize chemical structure, especially in fields such as nanotechnology where material properties are highly dependent on the size and shape of nanoparticles. Large data sources of nanoparticle information stemming from microscopy images would thus be highly beneficial. Millions of microscopy images exist, but they lie fragmented across the literature, typically presented individually within a paper article and usually in a qualitative fashion therein, even though they harbor a wealth of numeric information. We present the ImageDataExtractor toolkit that autoidentifies and autoextracts microscopy images from scientific documents, whereupon it autonomously analyzes each image to produce quantitative particle size and shape information about its subject material. Each image is quantified by decoding its scale bar information using optical character recognition, with help from super-resolution convolutional neural networks where required. Individual particles are detected and profiled using various thresholding, segmentation, polygon fitting, and edge correction routines. The high-throughput operational capability of ImageDataExtractor means that it can be used to generate large-data sources of particle information for data-driven materials discovery. Evaluation metrics, precision and recall, are greater than 80% for the majority of the image processing steps, and precision is above 80% for all critical steps. The ImageDataExtractor tool is released under the MIT license and is available to download from http://www.imagedataextractor.org.