On the use of Physics in Machine Learning for Manufacturing Process Inspection

George Barbastathis; Qihang Zhang; Ajinkya Pandit; Wenlong Tang; Charles Papageorgiou; Richard D. Braatz; Allan S. Myerson; Bingyao Tan; Leopold Schmetterer

doi:10.1117/12.2678259

On the use of Physics in Machine Learning for Manufacturing Process Inspection

George Barbastathis, Qihang Zhang, Ajinkya Pandit, Wenlong Tang^*, Charles Papageorgiou, Richard D. Braatz, Allan S. Myerson, Bingyao Tan^*, Leopold Schmetterer^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We discuss the use of machine learning in computational imaging for manufacturing process inspection and control. In a recent article¹ we described a physics-enhanced auto-correlation based estimator (Peace) for quantitative speckle. We derived an explicit forward relationship between the particle size distribution (PSD) and the speckle autocorrelation for particle sizes significantly larger than the wavelength (×100 „ ×1,000). We subsequently trained a machine learning kernel to invert the autocorrelation and obtain the PSD, using the explicit forward model to reduce the number of experimentally acquired examples. In this talk, we present an expanded discussion of Peace and its properties, including spatial and temporal sampling and accuracy, and more general applications.

Original language	English
Title of host publication	Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI
Editors	Pietro Ferraro, Demetri Psaltis, Simonetta Grilli
Publisher	SPIE
ISBN (Electronic)	9781510664531
DOIs	https://doi.org/10.1117/12.2678259
Publication status	Published - 2023
Externally published	Yes
Event	Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI 2023 - Munich, Germany Duration: Jun 26 2023 → Jun 29 2023

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	12622
ISSN (Print)	0277-786X
ISSN (Electronic)	1996-756X

Conference

Conference	Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI 2023
Country/Territory	Germany
City	Munich
Period	6/26/23 → 6/29/23

Bibliographical note

Publisher Copyright:
© 2023 SPIE. All rights reserved.

ASJC Scopus Subject Areas

Electronic, Optical and Magnetic Materials
Condensed Matter Physics
Computer Science Applications
Applied Mathematics
Electrical and Electronic Engineering

Access to Document

10.1117/12.2678259

Cite this

Barbastathis, G., Zhang, Q., Pandit, A., Tang, W., Papageorgiou, C., Braatz, R. D., Myerson, A. S., Tan, B., & Schmetterer, L. (2023). On the use of Physics in Machine Learning for Manufacturing Process Inspection. In P. Ferraro, D. Psaltis, & S. Grilli (Eds.), Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI Article 126220Y (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 12622). SPIE. https://doi.org/10.1117/12.2678259

Barbastathis, George ; Zhang, Qihang ; Pandit, Ajinkya et al. / On the use of Physics in Machine Learning for Manufacturing Process Inspection. Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI. editor / Pietro Ferraro ; Demetri Psaltis ; Simonetta Grilli. SPIE, 2023. (Proceedings of SPIE - The International Society for Optical Engineering).

@inproceedings{fe6a4830e7564c378fe32886467491fa,

title = "On the use of Physics in Machine Learning for Manufacturing Process Inspection",

abstract = "We discuss the use of machine learning in computational imaging for manufacturing process inspection and control. In a recent article1 we described a physics-enhanced auto-correlation based estimator (Peace) for quantitative speckle. We derived an explicit forward relationship between the particle size distribution (PSD) and the speckle autocorrelation for particle sizes significantly larger than the wavelength (×100 „ ×1,000). We subsequently trained a machine learning kernel to invert the autocorrelation and obtain the PSD, using the explicit forward model to reduce the number of experimentally acquired examples. In this talk, we present an expanded discussion of Peace and its properties, including spatial and temporal sampling and accuracy, and more general applications.",

author = "George Barbastathis and Qihang Zhang and Ajinkya Pandit and Wenlong Tang and Charles Papageorgiou and Braatz, {Richard D.} and Myerson, {Allan S.} and Bingyao Tan and Leopold Schmetterer",

note = "Publisher Copyright: {\textcopyright} 2023 SPIE. All rights reserved.; Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI 2023 ; Conference date: 26-06-2023 Through 29-06-2023",

year = "2023",

doi = "10.1117/12.2678259",

language = "English",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

publisher = "SPIE",

editor = "Pietro Ferraro and Demetri Psaltis and Simonetta Grilli",

booktitle = "Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI",

address = "United States",

}

Barbastathis, G, Zhang, Q, Pandit, A, Tang, W, Papageorgiou, C, Braatz, RD, Myerson, AS, Tan, B & Schmetterer, L 2023, On the use of Physics in Machine Learning for Manufacturing Process Inspection. in P Ferraro, D Psaltis & S Grilli (eds), Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI., 126220Y, Proceedings of SPIE - The International Society for Optical Engineering, vol. 12622, SPIE, Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI 2023, Munich, Germany, 6/26/23. https://doi.org/10.1117/12.2678259

On the use of Physics in Machine Learning for Manufacturing Process Inspection. / Barbastathis, George; Zhang, Qihang; Pandit, Ajinkya et al.
Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI. ed. / Pietro Ferraro; Demetri Psaltis; Simonetta Grilli. SPIE, 2023. 126220Y (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 12622).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - On the use of Physics in Machine Learning for Manufacturing Process Inspection

AU - Barbastathis, George

AU - Zhang, Qihang

AU - Pandit, Ajinkya

AU - Tang, Wenlong

AU - Papageorgiou, Charles

AU - Braatz, Richard D.

AU - Myerson, Allan S.

AU - Tan, Bingyao

AU - Schmetterer, Leopold

PY - 2023

Y1 - 2023

N2 - We discuss the use of machine learning in computational imaging for manufacturing process inspection and control. In a recent article1 we described a physics-enhanced auto-correlation based estimator (Peace) for quantitative speckle. We derived an explicit forward relationship between the particle size distribution (PSD) and the speckle autocorrelation for particle sizes significantly larger than the wavelength (×100 „ ×1,000). We subsequently trained a machine learning kernel to invert the autocorrelation and obtain the PSD, using the explicit forward model to reduce the number of experimentally acquired examples. In this talk, we present an expanded discussion of Peace and its properties, including spatial and temporal sampling and accuracy, and more general applications.

AB - We discuss the use of machine learning in computational imaging for manufacturing process inspection and control. In a recent article1 we described a physics-enhanced auto-correlation based estimator (Peace) for quantitative speckle. We derived an explicit forward relationship between the particle size distribution (PSD) and the speckle autocorrelation for particle sizes significantly larger than the wavelength (×100 „ ×1,000). We subsequently trained a machine learning kernel to invert the autocorrelation and obtain the PSD, using the explicit forward model to reduce the number of experimentally acquired examples. In this talk, we present an expanded discussion of Peace and its properties, including spatial and temporal sampling and accuracy, and more general applications.

UR - http://www.scopus.com/inward/record.url?scp=85172912822&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85172912822&partnerID=8YFLogxK

U2 - 10.1117/12.2678259

DO - 10.1117/12.2678259

M3 - Conference contribution

AN - SCOPUS:85172912822

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI

A2 - Ferraro, Pietro

A2 - Psaltis, Demetri

A2 - Grilli, Simonetta

PB - SPIE

T2 - Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI 2023

Y2 - 26 June 2023 through 29 June 2023

ER -

Barbastathis G, Zhang Q, Pandit A, Tang W, Papageorgiou C, Braatz RD et al. On the use of Physics in Machine Learning for Manufacturing Process Inspection. In Ferraro P, Psaltis D, Grilli S, editors, Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI. SPIE. 2023. 126220Y. (Proceedings of SPIE - The International Society for Optical Engineering). doi: 10.1117/12.2678259

On the use of Physics in Machine Learning for Manufacturing Process Inspection

Abstract

Publication series

Conference

Bibliographical note

ASJC Scopus Subject Areas

Access to Document

Other files and links

Fingerprint

Cite this