Trainable segmentation for transmission electron microscope images of inorganic nanoparticles

DOI: 10.1111/jmi.13110

Authors: Cameron G. Bell (Diamond Light Source) , Kevin P. Treder (University of Oxford) , Judy S. Kim (University of Oxford; The Rosalind Franklin Institute) , Manfred E. Schuster (Johnson Matthey Technology Centre) , Angus I. Kirkland (University of Oxford; Diamond Light Source; The Rosalind Franklin Institute) , Thomas J. A. Slater (Diamond Light Source; Cardiff University)
Co-authored by industrial partner: Yes

Type: Journal Paper
Journal: Journal Of Microscopy

State: Published (Approved)
Published: May 2022
Diamond Proposal Number(s): 26559

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.

Subject Areas: Technique Development, Materials, Information and Communication Technology

Diamond Offline Facilities: Electron Physical Sciences Imaging Centre (ePSIC)
Instruments: E01-JEM ARM 200CF

Added On: 04/05/2022 10:54

Discipline Tags:

Technique Development - Materials Science Computing & software technologies Information & Communication Technologies Materials Science Nanoscience/Nanotechnology

Technical Tags:

Microscopy Electron Microscopy (EM) Scanning Transmission Electron Microscopy (STEM)