A lightweight transformer for faster and robust EBSD data collection
Abstract Three dimensional electron back-scattered diffraction (EBSD) microscopy is a critical tool in many applications in materials science, yet its data quality can fluctuate greatly during the arduous collection process, particularly via serial-sectioning. Fortunately, 3D EBSD data is inherently...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-12-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-47936-6 |
| _version_ | 1797415272838070272 |
|---|---|
| author | Harry Dong Sean Donegan Megna Shah Yuejie Chi |
| author_facet | Harry Dong Sean Donegan Megna Shah Yuejie Chi |
| author_sort | Harry Dong |
| collection | DOAJ |
| description | Abstract Three dimensional electron back-scattered diffraction (EBSD) microscopy is a critical tool in many applications in materials science, yet its data quality can fluctuate greatly during the arduous collection process, particularly via serial-sectioning. Fortunately, 3D EBSD data is inherently sequential, opening up the opportunity to use transformers, state-of-the-art deep learning architectures that have made breakthroughs in a plethora of domains, for data processing and recovery. To be more robust to errors and accelerate this 3D EBSD data collection, we introduce a two step method that recovers missing slices in an 3D EBSD volume, using an efficient transformer model and a projection algorithm to process the transformer’s outputs. Overcoming the computational and practical hurdles of deep learning with scarce high dimensional data, we train this model using only synthetic 3D EBSD data with self-supervision and obtain superior recovery accuracy on real 3D EBSD data, compared to existing methods. |
| first_indexed | 2024-03-09T05:45:19Z |
| format | Article |
| id | doaj.art-a9dfd7baf3f94f54994c0050177ab59c |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-03-09T05:45:19Z |
| publishDate | 2023-12-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-a9dfd7baf3f94f54994c0050177ab59c2023-12-03T12:21:49ZengNature PortfolioScientific Reports2045-23222023-12-0113111310.1038/s41598-023-47936-6A lightweight transformer for faster and robust EBSD data collectionHarry Dong0Sean Donegan1Megna Shah2Yuejie Chi3Department of Electrical and Computer Engineering, Carnegie Mellon UniversityAir Force Research Laboratory, Materials and Manufacturing Directorate, Wright-Patterson AFBAir Force Research Laboratory, Materials and Manufacturing Directorate, Wright-Patterson AFBDepartment of Electrical and Computer Engineering, Carnegie Mellon UniversityAbstract Three dimensional electron back-scattered diffraction (EBSD) microscopy is a critical tool in many applications in materials science, yet its data quality can fluctuate greatly during the arduous collection process, particularly via serial-sectioning. Fortunately, 3D EBSD data is inherently sequential, opening up the opportunity to use transformers, state-of-the-art deep learning architectures that have made breakthroughs in a plethora of domains, for data processing and recovery. To be more robust to errors and accelerate this 3D EBSD data collection, we introduce a two step method that recovers missing slices in an 3D EBSD volume, using an efficient transformer model and a projection algorithm to process the transformer’s outputs. Overcoming the computational and practical hurdles of deep learning with scarce high dimensional data, we train this model using only synthetic 3D EBSD data with self-supervision and obtain superior recovery accuracy on real 3D EBSD data, compared to existing methods.https://doi.org/10.1038/s41598-023-47936-6 |
| spellingShingle | Harry Dong Sean Donegan Megna Shah Yuejie Chi A lightweight transformer for faster and robust EBSD data collection Scientific Reports |
| title | A lightweight transformer for faster and robust EBSD data collection |
| title_full | A lightweight transformer for faster and robust EBSD data collection |
| title_fullStr | A lightweight transformer for faster and robust EBSD data collection |
| title_full_unstemmed | A lightweight transformer for faster and robust EBSD data collection |
| title_short | A lightweight transformer for faster and robust EBSD data collection |
| title_sort | lightweight transformer for faster and robust ebsd data collection |
| url | https://doi.org/10.1038/s41598-023-47936-6 |
| work_keys_str_mv | AT harrydong alightweighttransformerforfasterandrobustebsddatacollection AT seandonegan alightweighttransformerforfasterandrobustebsddatacollection AT megnashah alightweighttransformerforfasterandrobustebsddatacollection AT yuejiechi alightweighttransformerforfasterandrobustebsddatacollection AT harrydong lightweighttransformerforfasterandrobustebsddatacollection AT seandonegan lightweighttransformerforfasterandrobustebsddatacollection AT megnashah lightweighttransformerforfasterandrobustebsddatacollection AT yuejiechi lightweighttransformerforfasterandrobustebsddatacollection |