AtomVision: A machine vision library for atomistic images

Abstract

Computer vision techniques have immense potential for materials design applications. In this work, we introduce an integrated and general-purpose AtomVision library that can be used to generate, curate scanning tunneling microscopy (STM) and scanning transmission electron microscopy (STEM) datasets and apply machine learning techniques. To demonstrate the applicability of this library, we 1) generate and curate an atomistic image dataset of about 10000 materials, 2) develop and compare convolutional and graph neural network models to classify the Bravais lattices, 3) develop fully convolutional neural network using U-Net architecture to pixelwise classify atom vs background, 4) use generative adversarial network for super-resolution, 5) curate a natural language processing based image dataset using open-access arXiv dataset, and 6) integrate the computational framework with experimental microscopy tools. AtomVision library is available at https://github.com/usnistgov/atomvision.

References in corpus (5)

• scikit-image: Image processing in Python
• Scanning Tunneling Spectroscopy of Graphene on Graphite
• EXSCLAIM! -- An automated pipeline for the construction of labeled materials imaging datasets from literature
• Effects of impurities on Tamm-like lanthanide-metal surface states
• Correlations between the morphology and the electronic structure at the surface of thin film manganites, investigated with STM

Cited by in corpus (5)

• JARVIS-Leaderboard: A Large Scale Benchmark of Materials Design Methods
• ChemNLP: A Natural Language Processing based Library for Materials Chemistry Text Data
• Recent progress in the JARVIS infrastructure for next-generation data-driven materials design
• InterMat: Accelerating Band Offset Prediction in Semiconductor Interfaces with DFT and Deep Learning
• The JARVIS Infrastructure is All You Need for Materials Design