Detecting Outliers with Poisson Image Interpolation

Jeremy Tan; Benjamin Hou; Thomas Day; John Simpson; Daniel Rueckert; Bernhard Kainz

doi:10.1007/978-3-030-87240-3_56

Detecting Outliers with Poisson Image Interpolation

Jeremy Tan^*, Benjamin Hou, Thomas Day, John Simpson, Daniel Rueckert, Bernhard Kainz

^*Corresponding author for this work

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

13 Citations (Scopus)

Abstract

Supervised learning of every possible pathology is unrealistic for many primary care applications like health screening. Image anomaly detection methods that learn normal appearance from only healthy data have shown promising results recently. We propose an alternative to image reconstruction-based and image embedding-based methods and propose a new self-supervised method to tackle pathological anomaly detection. Our approach originates in the foreign patch interpolation (FPI) strategy that has shown superior performance on brain MRI and abdominal CT data. We propose to use a better patch interpolation strategy, Poisson image interpolation (PII), which makes our method suitable for applications in challenging data regimes. PII outperforms state-of-the-art methods by a good margin when tested on surrogate tasks like identifying common lung anomalies in chest X-rays or hypo-plastic left heart syndrome in prenatal, fetal cardiac ultrasound images. Code available at https://github.com/jemtan/PII.

Original language	English
Title of host publication	Medical Image Computing and Computer Assisted Intervention – MICCAI 2021 - 24th International Conference, Proceedings
Editors	Marleen de Bruijne, Philippe C. Cattin, Stéphane Cotin, Nicolas Padoy, Stefanie Speidel, Yefeng Zheng, Caroline Essert
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	581-591
Number of pages	11
ISBN (Print)	9783030872397
DOIs	https://doi.org/10.1007/978-3-030-87240-3_56
Publication status	Published - 2021
Externally published	Yes
Event	24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021 - Virtual, Online Duration: 27 Sept 2021 → 1 Oct 2021

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12905 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021
City	Virtual, Online
Period	27/09/21 → 1/10/21

Bibliographical note

Funding Information:
Acknowledgements. Support from Wellcome Trust IEH Award iFind project [102431] and UK Research and Innovation London Medical Imaging and Artificial Intelligence Centre for Value Based Healthcare. JT was supported by the ICL President’s Scholarship.

Publisher Copyright:
© 2021, Springer Nature Switzerland AG.

Keywords

Outlier detection
Self-supervised learning

Access to Document

10.1007/978-3-030-87240-3_56

Cite this

Tan, J., Hou, B., Day, T., Simpson, J., Rueckert, D., & Kainz, B. (2021). Detecting Outliers with Poisson Image Interpolation. In M. de Bruijne, P. C. Cattin, S. Cotin, N. Padoy, S. Speidel, Y. Zheng, & C. Essert (Eds.), Medical Image Computing and Computer Assisted Intervention – MICCAI 2021 - 24th International Conference, Proceedings (pp. 581-591). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12905 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-87240-3_56

Tan, Jeremy ; Hou, Benjamin ; Day, Thomas et al. / Detecting Outliers with Poisson Image Interpolation. Medical Image Computing and Computer Assisted Intervention – MICCAI 2021 - 24th International Conference, Proceedings. editor / Marleen de Bruijne ; Philippe C. Cattin ; Stéphane Cotin ; Nicolas Padoy ; Stefanie Speidel ; Yefeng Zheng ; Caroline Essert. Springer Science and Business Media Deutschland GmbH, 2021. pp. 581-591 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{b796c560a8ca46739786b2c81243135c,

title = "Detecting Outliers with Poisson Image Interpolation",

abstract = "Supervised learning of every possible pathology is unrealistic for many primary care applications like health screening. Image anomaly detection methods that learn normal appearance from only healthy data have shown promising results recently. We propose an alternative to image reconstruction-based and image embedding-based methods and propose a new self-supervised method to tackle pathological anomaly detection. Our approach originates in the foreign patch interpolation (FPI) strategy that has shown superior performance on brain MRI and abdominal CT data. We propose to use a better patch interpolation strategy, Poisson image interpolation (PII), which makes our method suitable for applications in challenging data regimes. PII outperforms state-of-the-art methods by a good margin when tested on surrogate tasks like identifying common lung anomalies in chest X-rays or hypo-plastic left heart syndrome in prenatal, fetal cardiac ultrasound images. Code available at https://github.com/jemtan/PII.",

keywords = "Outlier detection, Self-supervised learning",

author = "Jeremy Tan and Benjamin Hou and Thomas Day and John Simpson and Daniel Rueckert and Bernhard Kainz",

note = "Funding Information: Acknowledgements. Support from Wellcome Trust IEH Award iFind project [102431] and UK Research and Innovation London Medical Imaging and Artificial Intelligence Centre for Value Based Healthcare. JT was supported by the ICL President{\textquoteright}s Scholarship. Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021 ; Conference date: 27-09-2021 Through 01-10-2021",

year = "2021",

doi = "10.1007/978-3-030-87240-3_56",

language = "English",

isbn = "9783030872397",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "581--591",

editor = "{de Bruijne}, Marleen and Cattin, {Philippe C.} and St{\'e}phane Cotin and Nicolas Padoy and Stefanie Speidel and Yefeng Zheng and Caroline Essert",

booktitle = "Medical Image Computing and Computer Assisted Intervention – MICCAI 2021 - 24th International Conference, Proceedings",

}

Tan, J, Hou, B, Day, T, Simpson, J, Rueckert, D & Kainz, B 2021, Detecting Outliers with Poisson Image Interpolation. in M de Bruijne, PC Cattin, S Cotin, N Padoy, S Speidel, Y Zheng & C Essert (eds), Medical Image Computing and Computer Assisted Intervention – MICCAI 2021 - 24th International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12905 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 581-591, 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021, Virtual, Online, 27/09/21. https://doi.org/10.1007/978-3-030-87240-3_56

Detecting Outliers with Poisson Image Interpolation. / Tan, Jeremy; Hou, Benjamin; Day, Thomas et al.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2021 - 24th International Conference, Proceedings. ed. / Marleen de Bruijne; Philippe C. Cattin; Stéphane Cotin; Nicolas Padoy; Stefanie Speidel; Yefeng Zheng; Caroline Essert. Springer Science and Business Media Deutschland GmbH, 2021. p. 581-591 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12905 LNCS).

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

TY - GEN

T1 - Detecting Outliers with Poisson Image Interpolation

AU - Tan, Jeremy

AU - Hou, Benjamin

AU - Day, Thomas

AU - Simpson, John

AU - Rueckert, Daniel

AU - Kainz, Bernhard

N1 - Funding Information: Acknowledgements. Support from Wellcome Trust IEH Award iFind project [102431] and UK Research and Innovation London Medical Imaging and Artificial Intelligence Centre for Value Based Healthcare. JT was supported by the ICL President’s Scholarship. Publisher Copyright: © 2021, Springer Nature Switzerland AG.

PY - 2021

Y1 - 2021

N2 - Supervised learning of every possible pathology is unrealistic for many primary care applications like health screening. Image anomaly detection methods that learn normal appearance from only healthy data have shown promising results recently. We propose an alternative to image reconstruction-based and image embedding-based methods and propose a new self-supervised method to tackle pathological anomaly detection. Our approach originates in the foreign patch interpolation (FPI) strategy that has shown superior performance on brain MRI and abdominal CT data. We propose to use a better patch interpolation strategy, Poisson image interpolation (PII), which makes our method suitable for applications in challenging data regimes. PII outperforms state-of-the-art methods by a good margin when tested on surrogate tasks like identifying common lung anomalies in chest X-rays or hypo-plastic left heart syndrome in prenatal, fetal cardiac ultrasound images. Code available at https://github.com/jemtan/PII.

AB - Supervised learning of every possible pathology is unrealistic for many primary care applications like health screening. Image anomaly detection methods that learn normal appearance from only healthy data have shown promising results recently. We propose an alternative to image reconstruction-based and image embedding-based methods and propose a new self-supervised method to tackle pathological anomaly detection. Our approach originates in the foreign patch interpolation (FPI) strategy that has shown superior performance on brain MRI and abdominal CT data. We propose to use a better patch interpolation strategy, Poisson image interpolation (PII), which makes our method suitable for applications in challenging data regimes. PII outperforms state-of-the-art methods by a good margin when tested on surrogate tasks like identifying common lung anomalies in chest X-rays or hypo-plastic left heart syndrome in prenatal, fetal cardiac ultrasound images. Code available at https://github.com/jemtan/PII.

KW - Outlier detection

KW - Self-supervised learning

UR - http://www.scopus.com/inward/record.url?scp=85115388153&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-87240-3_56

DO - 10.1007/978-3-030-87240-3_56

M3 - Conference contribution

AN - SCOPUS:85115388153

SN - 9783030872397

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 581

EP - 591

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2021 - 24th International Conference, Proceedings

A2 - de Bruijne, Marleen

A2 - Cattin, Philippe C.

A2 - Cotin, Stéphane

A2 - Padoy, Nicolas

A2 - Speidel, Stefanie

A2 - Zheng, Yefeng

A2 - Essert, Caroline

PB - Springer Science and Business Media Deutschland GmbH

T2 - 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021

Y2 - 27 September 2021 through 1 October 2021

ER -

Tan J, Hou B, Day T, Simpson J, Rueckert D, Kainz B. Detecting Outliers with Poisson Image Interpolation. In de Bruijne M, Cattin PC, Cotin S, Padoy N, Speidel S, Zheng Y, Essert C, editors, Medical Image Computing and Computer Assisted Intervention – MICCAI 2021 - 24th International Conference, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 581-591. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-87240-3_56

Detecting Outliers with Poisson Image Interpolation

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this