Weakly Semi-supervised Detection in Lung Ultrasound Videos

Jiahong Ouyang; Li Chen; Gary Y. Li; Naveen Balaraju; Shubham Patil; Courosh Mehanian; Sourabh Kulhare; Rachel Millin; Kenton W. Gregory; Cynthia R. Gregory; Meihua Zhu; David O. Kessler; Laurie Malia; Almaz Dessie; Joni Rabiner; Di Coneybeare; Bo Shopsin; Andrew Hersh; Cristian Madar; Jeffrey Shupp; Laura S. Johnson; Jacob Avila; Kristin Dwyer; Peter Weimersheimer; Balasundar Raju; Jochen Kruecker; Alvin Chen

doi:10.1007/978-3-031-34048-2_16

Weakly Semi-supervised Detection in Lung Ultrasound Videos

Jiahong Ouyang, Li Chen, Gary Y. Li, Naveen Balaraju, Shubham Patil, Courosh Mehanian, Sourabh Kulhare, Rachel Millin, Kenton W. Gregory, Cynthia R. Gregory, Meihua Zhu, David O. Kessler, Laurie Malia, Almaz Dessie, Joni Rabiner, Di Coneybeare, Bo Shopsin, Andrew Hersh, Cristian Madar, Jeffrey ShuppLaura S. Johnson, Jacob Avila, Kristin Dwyer, Peter Weimersheimer, Balasundar Raju, Jochen Kruecker, Alvin Chen

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

2 Scopus citations

Abstract

Frame-by-frame annotation of bounding boxes by clinical experts is often required to train fully supervised object detection models on medical video data. We propose a method for improving object detection in medical videos through weak supervision from video-level labels. More concretely, we aggregate individual detection predictions into video-level predictions and extend a teacher-student training strategy to provide additional supervision via a video-level loss. We also introduce improvements to the underlying teacher-student framework, including methods to improve the quality of pseudo-labels based on weak supervision and adaptive schemes to optimize knowledge transfer between the student and teacher networks. We apply this approach to the clinically important task of detecting lung consolidations (seen in respiratory infections such as COVID-19 pneumonia) in medical ultrasound videos. Experiments reveal that our framework improves detection accuracy and robustness compared to baseline semi-supervised models, and improves efficiency in data and annotation usage.

Original language	English (US)
Title of host publication	Information Processing in Medical Imaging - 28th International Conference, IPMI 2023, Proceedings
Editors	Alejandro Frangi, Marleen de Bruijne, Demian Wassermann, Nassir Navab
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	195-207
Number of pages	13
ISBN (Print)	9783031340475
DOIs	https://doi.org/10.1007/978-3-031-34048-2_16
State	Published - 2023
Event	28th International Conference on Information Processing in Medical Imaging, IPMI 2023 - San Carlos de Bariloche, Argentina Duration: Jun 18 2023 → Jun 23 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13939 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	28th International Conference on Information Processing in Medical Imaging, IPMI 2023
Country/Territory	Argentina
City	San Carlos de Bariloche
Period	6/18/23 → 6/23/23

Keywords

Medical Ultrasound
Object Detection
Semi-Supervised Learning
Weakly Supervised Learning

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-34048-2_16

Cite this

Ouyang, J., Chen, L., Li, G. Y., Balaraju, N., Patil, S., Mehanian, C., Kulhare, S., Millin, R., Gregory, K. W., Gregory, C. R., Zhu, M., Kessler, D. O., Malia, L., Dessie, A., Rabiner, J., Coneybeare, D., Shopsin, B., Hersh, A., Madar, C., ... Chen, A. (2023). Weakly Semi-supervised Detection in Lung Ultrasound Videos. In A. Frangi, M. de Bruijne, D. Wassermann, & N. Navab (Eds.), Information Processing in Medical Imaging - 28th International Conference, IPMI 2023, Proceedings (pp. 195-207). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13939 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-34048-2_16

Weakly Semi-supervised Detection in Lung Ultrasound Videos. / Ouyang, Jiahong; Chen, Li; Li, Gary Y. et al.
Information Processing in Medical Imaging - 28th International Conference, IPMI 2023, Proceedings. ed. / Alejandro Frangi; Marleen de Bruijne; Demian Wassermann; Nassir Navab. Springer Science and Business Media Deutschland GmbH, 2023. p. 195-207 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13939 LNCS).

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

Ouyang, J, Chen, L, Li, GY, Balaraju, N, Patil, S, Mehanian, C, Kulhare, S, Millin, R, Gregory, KW , Gregory, CR, Zhu, M, Kessler, DO, Malia, L, Dessie, A, Rabiner, J, Coneybeare, D, Shopsin, B, Hersh, A, Madar, C, Shupp, J, Johnson, LS, Avila, J, Dwyer, K, Weimersheimer, P, Raju, B, Kruecker, J & Chen, A 2023, Weakly Semi-supervised Detection in Lung Ultrasound Videos. in A Frangi, M de Bruijne, D Wassermann & N Navab (eds), Information Processing in Medical Imaging - 28th International Conference, IPMI 2023, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13939 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 195-207, 28th International Conference on Information Processing in Medical Imaging, IPMI 2023, San Carlos de Bariloche, Argentina, 6/18/23. https://doi.org/10.1007/978-3-031-34048-2_16

Ouyang J, Chen L, Li GY, Balaraju N, Patil S, Mehanian C et al. Weakly Semi-supervised Detection in Lung Ultrasound Videos. In Frangi A, de Bruijne M, Wassermann D, Navab N, editors, Information Processing in Medical Imaging - 28th International Conference, IPMI 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 195-207. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-34048-2_16

Ouyang, Jiahong ; Chen, Li ; Li, Gary Y. et al. / Weakly Semi-supervised Detection in Lung Ultrasound Videos. Information Processing in Medical Imaging - 28th International Conference, IPMI 2023, Proceedings. editor / Alejandro Frangi ; Marleen de Bruijne ; Demian Wassermann ; Nassir Navab. Springer Science and Business Media Deutschland GmbH, 2023. pp. 195-207 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{75e90bee22754cb0beae084c846ef5f0,

title = "Weakly Semi-supervised Detection in Lung Ultrasound Videos",

abstract = "Frame-by-frame annotation of bounding boxes by clinical experts is often required to train fully supervised object detection models on medical video data. We propose a method for improving object detection in medical videos through weak supervision from video-level labels. More concretely, we aggregate individual detection predictions into video-level predictions and extend a teacher-student training strategy to provide additional supervision via a video-level loss. We also introduce improvements to the underlying teacher-student framework, including methods to improve the quality of pseudo-labels based on weak supervision and adaptive schemes to optimize knowledge transfer between the student and teacher networks. We apply this approach to the clinically important task of detecting lung consolidations (seen in respiratory infections such as COVID-19 pneumonia) in medical ultrasound videos. Experiments reveal that our framework improves detection accuracy and robustness compared to baseline semi-supervised models, and improves efficiency in data and annotation usage.",

keywords = "Medical Ultrasound, Object Detection, Semi-Supervised Learning, Weakly Supervised Learning",

author = "Jiahong Ouyang and Li Chen and Li, {Gary Y.} and Naveen Balaraju and Shubham Patil and Courosh Mehanian and Sourabh Kulhare and Rachel Millin and Gregory, {Kenton W.} and Gregory, {Cynthia R.} and Meihua Zhu and Kessler, {David O.} and Laurie Malia and Almaz Dessie and Joni Rabiner and Di Coneybeare and Bo Shopsin and Andrew Hersh and Cristian Madar and Jeffrey Shupp and Johnson, {Laura S.} and Jacob Avila and Kristin Dwyer and Peter Weimersheimer and Balasundar Raju and Jochen Kruecker and Alvin Chen",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 28th International Conference on Information Processing in Medical Imaging, IPMI 2023 ; Conference date: 18-06-2023 Through 23-06-2023",

year = "2023",

doi = "10.1007/978-3-031-34048-2_16",

language = "English (US)",

isbn = "9783031340475",

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 = "195--207",

editor = "Alejandro Frangi and {de Bruijne}, Marleen and Demian Wassermann and Nassir Navab",

booktitle = "Information Processing in Medical Imaging - 28th International Conference, IPMI 2023, Proceedings",

address = "Germany",

}

TY - GEN

T1 - Weakly Semi-supervised Detection in Lung Ultrasound Videos

AU - Ouyang, Jiahong

AU - Chen, Li

AU - Li, Gary Y.

AU - Balaraju, Naveen

AU - Patil, Shubham

AU - Mehanian, Courosh

AU - Kulhare, Sourabh

AU - Millin, Rachel

AU - Gregory, Kenton W.

AU - Gregory, Cynthia R.

AU - Zhu, Meihua

AU - Kessler, David O.

AU - Malia, Laurie

AU - Dessie, Almaz

AU - Rabiner, Joni

AU - Coneybeare, Di

AU - Shopsin, Bo

AU - Hersh, Andrew

AU - Madar, Cristian

AU - Shupp, Jeffrey

AU - Johnson, Laura S.

AU - Avila, Jacob

AU - Dwyer, Kristin

AU - Weimersheimer, Peter

AU - Raju, Balasundar

AU - Kruecker, Jochen

AU - Chen, Alvin

PY - 2023

Y1 - 2023

N2 - Frame-by-frame annotation of bounding boxes by clinical experts is often required to train fully supervised object detection models on medical video data. We propose a method for improving object detection in medical videos through weak supervision from video-level labels. More concretely, we aggregate individual detection predictions into video-level predictions and extend a teacher-student training strategy to provide additional supervision via a video-level loss. We also introduce improvements to the underlying teacher-student framework, including methods to improve the quality of pseudo-labels based on weak supervision and adaptive schemes to optimize knowledge transfer between the student and teacher networks. We apply this approach to the clinically important task of detecting lung consolidations (seen in respiratory infections such as COVID-19 pneumonia) in medical ultrasound videos. Experiments reveal that our framework improves detection accuracy and robustness compared to baseline semi-supervised models, and improves efficiency in data and annotation usage.

AB - Frame-by-frame annotation of bounding boxes by clinical experts is often required to train fully supervised object detection models on medical video data. We propose a method for improving object detection in medical videos through weak supervision from video-level labels. More concretely, we aggregate individual detection predictions into video-level predictions and extend a teacher-student training strategy to provide additional supervision via a video-level loss. We also introduce improvements to the underlying teacher-student framework, including methods to improve the quality of pseudo-labels based on weak supervision and adaptive schemes to optimize knowledge transfer between the student and teacher networks. We apply this approach to the clinically important task of detecting lung consolidations (seen in respiratory infections such as COVID-19 pneumonia) in medical ultrasound videos. Experiments reveal that our framework improves detection accuracy and robustness compared to baseline semi-supervised models, and improves efficiency in data and annotation usage.

KW - Medical Ultrasound

KW - Object Detection

KW - Semi-Supervised Learning

KW - Weakly Supervised Learning

UR - http://www.scopus.com/inward/record.url?scp=85164022304&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85164022304&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-34048-2_16

DO - 10.1007/978-3-031-34048-2_16

M3 - Conference contribution

AN - SCOPUS:85164022304

SN - 9783031340475

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

SP - 195

EP - 207

BT - Information Processing in Medical Imaging - 28th International Conference, IPMI 2023, Proceedings

A2 - Frangi, Alejandro

A2 - de Bruijne, Marleen

A2 - Wassermann, Demian

A2 - Navab, Nassir

PB - Springer Science and Business Media Deutschland GmbH

T2 - 28th International Conference on Information Processing in Medical Imaging, IPMI 2023

Y2 - 18 June 2023 through 23 June 2023

ER -