Preprints
https://doi.org/10.5194/gchron-2022-12
https://doi.org/10.5194/gchron-2022-12
 
05 May 2022
05 May 2022
Status: this preprint is currently under review for the journal GChron.

Technical Note: colab_zirc_dims: a Google-Colab-based Toolset for Automated and Semi-automated Measurement of Mineral Grains in LA-ICP-MS Images Using Deep Learning Models

Michael C. Sitar1 and Ryan J. Leary2 Michael C. Sitar and Ryan J. Leary
  • 1Department of Geosciences, Colorado State University, Fort Collins, CO, 80523-1482, USA
  • 2Department of Earth and Environmental Science, New Mexico Institute of Mining and Technology, Socorro, NM, 87801, USA

Abstract. Collecting grain measurements for large detrital zircon age datasets is time-consuming, but a growing number of studies suggest such data are essential to understanding complex roles of grain size and morphology in grain transport and as indicators for grain provenance. We developed the colab_zirc_dims Python package to automate deep-learning-based segmentation and measurement of mineral grains from scaled images captured during laser ablation at facilities that use Chromium targeting software. The colab_zirc_dims package is implemented in a collection of freely accessible, ready-to-run Google Colab notebooks with additional functionalities for dataset preparation and semi-automated grain segmentation and measurement using a simple graphical user interface. Our automated grain measurement algorithm approaches human measurement accuracy when applied to a manually measured n = 5,004 detrital zircon dataset, but persistent errors necessitate semi-automated measurement for production of publication-quality datasets. We estimate that our semi-automated grain segmentation workflow will enable users to collect grain measurements for large (n ≥ 5,000), applicable datasets in under a day of work, and we hope that the colab_zirc_dims toolset allows more researchers to augment their detrital geochronology datasets with grain measurements.

Michael C. Sitar and Ryan J. Leary

Status: final response (author comments only)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on gchron-2022-12', Simon Nachtergaele, 01 Jun 2022
    • AC1: 'Reply on RC1', Michael Sitar, 22 Jul 2022
  • RC2: 'Comment on gchron-2022-12', Taryn Scharf, 12 Jun 2022
    • AC2: 'Reply on RC2', Michael Sitar, 22 Jul 2022

Michael C. Sitar and Ryan J. Leary

Data sets

Results, datasets, and replication code repository Michael C. Sitar and Ryan J. Leary https://doi.org/10.5281/zenodo.6412303

Model code and software

colab_zirc_dims v1.0.8 source code repository with archived links Michael C. Sitar https://doi.org/10.5281/ZENODO.6410745

Video supplement

colab_zirc_dims v1.0.8 Video Tutorial & Demo Michael C. Sitar https://youtu.be/WM7qEjaJdgo

Michael C. Sitar and Ryan J. Leary

Viewed

Total article views: 476 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
327 134 15 476 6 5
  • HTML: 327
  • PDF: 134
  • XML: 15
  • Total: 476
  • BibTeX: 6
  • EndNote: 5
Views and downloads (calculated since 05 May 2022)
Cumulative views and downloads (calculated since 05 May 2022)

Viewed (geographical distribution)

Total article views: 433 (including HTML, PDF, and XML) Thereof 433 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 22 Sep 2022
Download
Short summary
We developed code to automatically and semi-automatically measure dimensions of detrital mineral grains in reflected light images saved at LA-ICP-MS facilities that use Chromium targeting software. Our code uses trained deep learning models to segment grain-images with greater accuracy than is possible using other image segmentation techniques. We implement our code in user-accessible, ready-to-run Google Colab notebooks, links to which can be found at https://github.com/MCSitar/colab_zirc_dims.