Weakly Supervised Deep Metric Learning for Template Matching

Davit Buniatyan, Sergiy Popovych, Dodam Ih, Thomas Macrina, Jonathan Zung, Hyunjune Sebastian Seung

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Template matching by normalized cross correlation (NCC) is widely used for finding image correspondences. NCCNet improves the robustness of this algorithm by transforming image features with siamese convolutional nets trained to maximize the contrast between NCC values of true and false matches. The main technical contribution is a weakly supervised learning algorithm for the training. Unlike fully supervised approaches to metric learning, the method can improve upon vanilla NCC without receiving locations of true matches during training. The improvement is quantified through patches of brain images from serial section electron microscopy. Relative to a parameter-tuned bandpass filter, siamese convolutional nets significantly reduce false matches. The improved accuracy of the method could be essential for connectomics, because emerging petascale datasets may require billions of template matches during assembly. Our method is also expected to generalize to other computer vision applications that use template matching to find image correspondences.

Original languageEnglish (US)
Title of host publicationAdvances in Computer Vision - Proceedings of the 2019 Computer Vision Conference CVC
EditorsSupriya Kapoor, Kohei Arai
PublisherSpringer Verlag
Pages39-58
Number of pages20
ISBN (Print)9783030177942
DOIs
StatePublished - Jan 1 2020
EventComputer Vision Conference, CVC 2019 - Las Vegas, United States
Duration: Apr 25 2019Apr 26 2019

Publication series

NameAdvances in Intelligent Systems and Computing
Volume943

Conference

ConferenceComputer Vision Conference, CVC 2019
CountryUnited States
CityLas Vegas
Period4/25/194/26/19

Fingerprint

Template matching
Supervised learning
Bandpass filters
Learning algorithms
Computer vision
Electron microscopy
Brain

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Computer Science(all)

Cite this

Buniatyan, D., Popovych, S., Ih, D., Macrina, T., Zung, J., & Seung, H. S. (2020). Weakly Supervised Deep Metric Learning for Template Matching. In S. Kapoor, & K. Arai (Eds.), Advances in Computer Vision - Proceedings of the 2019 Computer Vision Conference CVC (pp. 39-58). (Advances in Intelligent Systems and Computing; Vol. 943). Springer Verlag. https://doi.org/10.1007/978-3-030-17795-9_4
Buniatyan, Davit ; Popovych, Sergiy ; Ih, Dodam ; Macrina, Thomas ; Zung, Jonathan ; Seung, Hyunjune Sebastian. / Weakly Supervised Deep Metric Learning for Template Matching. Advances in Computer Vision - Proceedings of the 2019 Computer Vision Conference CVC. editor / Supriya Kapoor ; Kohei Arai. Springer Verlag, 2020. pp. 39-58 (Advances in Intelligent Systems and Computing).
@inproceedings{4eb3f846c8d247848eb0c03b42964913,
title = "Weakly Supervised Deep Metric Learning for Template Matching",
abstract = "Template matching by normalized cross correlation (NCC) is widely used for finding image correspondences. NCCNet improves the robustness of this algorithm by transforming image features with siamese convolutional nets trained to maximize the contrast between NCC values of true and false matches. The main technical contribution is a weakly supervised learning algorithm for the training. Unlike fully supervised approaches to metric learning, the method can improve upon vanilla NCC without receiving locations of true matches during training. The improvement is quantified through patches of brain images from serial section electron microscopy. Relative to a parameter-tuned bandpass filter, siamese convolutional nets significantly reduce false matches. The improved accuracy of the method could be essential for connectomics, because emerging petascale datasets may require billions of template matches during assembly. Our method is also expected to generalize to other computer vision applications that use template matching to find image correspondences.",
author = "Davit Buniatyan and Sergiy Popovych and Dodam Ih and Thomas Macrina and Jonathan Zung and Seung, {Hyunjune Sebastian}",
year = "2020",
month = "1",
day = "1",
doi = "https://doi.org/10.1007/978-3-030-17795-9_4",
language = "English (US)",
isbn = "9783030177942",
series = "Advances in Intelligent Systems and Computing",
publisher = "Springer Verlag",
pages = "39--58",
editor = "Supriya Kapoor and Kohei Arai",
booktitle = "Advances in Computer Vision - Proceedings of the 2019 Computer Vision Conference CVC",
address = "Germany",

}

Buniatyan, D, Popovych, S, Ih, D, Macrina, T, Zung, J & Seung, HS 2020, Weakly Supervised Deep Metric Learning for Template Matching. in S Kapoor & K Arai (eds), Advances in Computer Vision - Proceedings of the 2019 Computer Vision Conference CVC. Advances in Intelligent Systems and Computing, vol. 943, Springer Verlag, pp. 39-58, Computer Vision Conference, CVC 2019, Las Vegas, United States, 4/25/19. https://doi.org/10.1007/978-3-030-17795-9_4

Weakly Supervised Deep Metric Learning for Template Matching. / Buniatyan, Davit; Popovych, Sergiy; Ih, Dodam; Macrina, Thomas; Zung, Jonathan; Seung, Hyunjune Sebastian.

Advances in Computer Vision - Proceedings of the 2019 Computer Vision Conference CVC. ed. / Supriya Kapoor; Kohei Arai. Springer Verlag, 2020. p. 39-58 (Advances in Intelligent Systems and Computing; Vol. 943).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Weakly Supervised Deep Metric Learning for Template Matching

AU - Buniatyan, Davit

AU - Popovych, Sergiy

AU - Ih, Dodam

AU - Macrina, Thomas

AU - Zung, Jonathan

AU - Seung, Hyunjune Sebastian

PY - 2020/1/1

Y1 - 2020/1/1

N2 - Template matching by normalized cross correlation (NCC) is widely used for finding image correspondences. NCCNet improves the robustness of this algorithm by transforming image features with siamese convolutional nets trained to maximize the contrast between NCC values of true and false matches. The main technical contribution is a weakly supervised learning algorithm for the training. Unlike fully supervised approaches to metric learning, the method can improve upon vanilla NCC without receiving locations of true matches during training. The improvement is quantified through patches of brain images from serial section electron microscopy. Relative to a parameter-tuned bandpass filter, siamese convolutional nets significantly reduce false matches. The improved accuracy of the method could be essential for connectomics, because emerging petascale datasets may require billions of template matches during assembly. Our method is also expected to generalize to other computer vision applications that use template matching to find image correspondences.

AB - Template matching by normalized cross correlation (NCC) is widely used for finding image correspondences. NCCNet improves the robustness of this algorithm by transforming image features with siamese convolutional nets trained to maximize the contrast between NCC values of true and false matches. The main technical contribution is a weakly supervised learning algorithm for the training. Unlike fully supervised approaches to metric learning, the method can improve upon vanilla NCC without receiving locations of true matches during training. The improvement is quantified through patches of brain images from serial section electron microscopy. Relative to a parameter-tuned bandpass filter, siamese convolutional nets significantly reduce false matches. The improved accuracy of the method could be essential for connectomics, because emerging petascale datasets may require billions of template matches during assembly. Our method is also expected to generalize to other computer vision applications that use template matching to find image correspondences.

UR - http://www.scopus.com/inward/record.url?scp=85065464426&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85065464426&partnerID=8YFLogxK

U2 - https://doi.org/10.1007/978-3-030-17795-9_4

DO - https://doi.org/10.1007/978-3-030-17795-9_4

M3 - Conference contribution

SN - 9783030177942

T3 - Advances in Intelligent Systems and Computing

SP - 39

EP - 58

BT - Advances in Computer Vision - Proceedings of the 2019 Computer Vision Conference CVC

A2 - Kapoor, Supriya

A2 - Arai, Kohei

PB - Springer Verlag

ER -

Buniatyan D, Popovych S, Ih D, Macrina T, Zung J, Seung HS. Weakly Supervised Deep Metric Learning for Template Matching. In Kapoor S, Arai K, editors, Advances in Computer Vision - Proceedings of the 2019 Computer Vision Conference CVC. Springer Verlag. 2020. p. 39-58. (Advances in Intelligent Systems and Computing). https://doi.org/10.1007/978-3-030-17795-9_4