The Trauma THOMPSON Challenge Report MICCAI 2023

Yupeng Zhuo; Andrew W. Kirkpatrick; Kyle Couperus; Oanh Tran; Juan Wachs

doi:10.1007/978-3-031-71626-3_8

The Trauma THOMPSON Challenge Report MICCAI 2023

Yupeng Zhuo, Andrew W. Kirkpatrick, Kyle Couperus, Oanh Tran, Juan Wachs

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

Abstract

Trauma TeleHelper for Operational Medical Procedure Support and Offline Network (Trauma THOMPSON) Challenge offers a first egocentric view dataset on emergency care procedures under resource constrained scenarios. This paper describes the baseline solutions to the four tasks of the Trauma THOMPSON Challenge 2023. They were not provided to the participants ahead of the challenge and was not part of the competition. The Temporal Segmentation Network algorithm is used for the action recognition task with top-1 accuracy of 11.38%. It is also used for the action anticipation task with top-1 accuracy of 5.82% and procedure recognition task with top-1 accuracy of 53.57%. ViLT is adopted for the Visual Question Answering challenge with an aggregate accuracy of 39.13%.

Original language	English
Title of host publication	AI for Brain Lesion Detection and Trauma Video Action Recognition - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge, Held in Conjunction with MICCAI 2023, Proceedings
Editors	Rina Bao, Ellen Grant, Yangming Ou, Andrew Kirkpatrick, Juan Wachs
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	61-71
Number of pages	11
ISBN (Print)	9783031716256
DOIs	https://doi.org/10.1007/978-3-031-71626-3_8
State	Published - 1 Jan 2025
Externally published	Yes
Event	1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge Held in Conjunction with 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023 - Vancouver, Canada Duration: 12 Oct 2023 → 16 Oct 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14567 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge Held in Conjunction with 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023
Country/Territory	Canada
City	Vancouver
Period	12/10/23 → 16/10/23

Keywords

Action Anticipation
Action Recognition
Procedure Recognition
Visual Question Answering

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-71626-3_8

Cite this

Zhuo, Y., W. Kirkpatrick, A., Couperus, K., Tran, O., & Wachs, J. (2025). The Trauma THOMPSON Challenge Report MICCAI 2023. In R. Bao, E. Grant, Y. Ou, A. Kirkpatrick, & J. Wachs (Eds.), AI for Brain Lesion Detection and Trauma Video Action Recognition - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge, Held in Conjunction with MICCAI 2023, Proceedings (pp. 61-71). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14567 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-71626-3_8

Zhuo, Yupeng ; W. Kirkpatrick, Andrew ; Couperus, Kyle et al. / The Trauma THOMPSON Challenge Report MICCAI 2023. AI for Brain Lesion Detection and Trauma Video Action Recognition - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge, Held in Conjunction with MICCAI 2023, Proceedings. editor / Rina Bao ; Ellen Grant ; Yangming Ou ; Andrew Kirkpatrick ; Juan Wachs. Springer Science and Business Media Deutschland GmbH, 2025. pp. 61-71 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{991e23873dac4f2a876066690941614b,

title = "The Trauma THOMPSON Challenge Report MICCAI 2023",

abstract = "Trauma TeleHelper for Operational Medical Procedure Support and Offline Network (Trauma THOMPSON) Challenge offers a first egocentric view dataset on emergency care procedures under resource constrained scenarios. This paper describes the baseline solutions to the four tasks of the Trauma THOMPSON Challenge 2023. They were not provided to the participants ahead of the challenge and was not part of the competition. The Temporal Segmentation Network algorithm is used for the action recognition task with top-1 accuracy of 11.38%. It is also used for the action anticipation task with top-1 accuracy of 5.82% and procedure recognition task with top-1 accuracy of 53.57%. ViLT is adopted for the Visual Question Answering challenge with an aggregate accuracy of 39.13%.",

keywords = "Action Anticipation, Action Recognition, Procedure Recognition, Visual Question Answering",

author = "Yupeng Zhuo and {W. Kirkpatrick}, Andrew and Kyle Couperus and Oanh Tran and Juan Wachs",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.; 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge Held in Conjunction with 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023 ; Conference date: 12-10-2023 Through 16-10-2023",

year = "2025",

month = jan,

day = "1",

doi = "10.1007/978-3-031-71626-3_8",

language = "English",

isbn = "9783031716256",

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 = "61--71",

editor = "Rina Bao and Ellen Grant and Yangming Ou and Andrew Kirkpatrick and Juan Wachs",

booktitle = "AI for Brain Lesion Detection and Trauma Video Action Recognition - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge, Held in Conjunction with MICCAI 2023, Proceedings",

address = "Germany",

}

Zhuo, Y, W. Kirkpatrick, A, Couperus, K, Tran, O & Wachs, J 2025, The Trauma THOMPSON Challenge Report MICCAI 2023. in R Bao, E Grant, Y Ou, A Kirkpatrick & J Wachs (eds), AI for Brain Lesion Detection and Trauma Video Action Recognition - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge, Held in Conjunction with MICCAI 2023, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14567 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 61-71, 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge Held in Conjunction with 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023, Vancouver, Canada, 12/10/23. https://doi.org/10.1007/978-3-031-71626-3_8

The Trauma THOMPSON Challenge Report MICCAI 2023. / Zhuo, Yupeng; W. Kirkpatrick, Andrew; Couperus, Kyle et al.
AI for Brain Lesion Detection and Trauma Video Action Recognition - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge, Held in Conjunction with MICCAI 2023, Proceedings. ed. / Rina Bao; Ellen Grant; Yangming Ou; Andrew Kirkpatrick; Juan Wachs. Springer Science and Business Media Deutschland GmbH, 2025. p. 61-71 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14567 LNCS).

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

TY - GEN

T1 - The Trauma THOMPSON Challenge Report MICCAI 2023

AU - Zhuo, Yupeng

AU - W. Kirkpatrick, Andrew

AU - Couperus, Kyle

AU - Tran, Oanh

AU - Wachs, Juan

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

PY - 2025/1/1

Y1 - 2025/1/1

N2 - Trauma TeleHelper for Operational Medical Procedure Support and Offline Network (Trauma THOMPSON) Challenge offers a first egocentric view dataset on emergency care procedures under resource constrained scenarios. This paper describes the baseline solutions to the four tasks of the Trauma THOMPSON Challenge 2023. They were not provided to the participants ahead of the challenge and was not part of the competition. The Temporal Segmentation Network algorithm is used for the action recognition task with top-1 accuracy of 11.38%. It is also used for the action anticipation task with top-1 accuracy of 5.82% and procedure recognition task with top-1 accuracy of 53.57%. ViLT is adopted for the Visual Question Answering challenge with an aggregate accuracy of 39.13%.

AB - Trauma TeleHelper for Operational Medical Procedure Support and Offline Network (Trauma THOMPSON) Challenge offers a first egocentric view dataset on emergency care procedures under resource constrained scenarios. This paper describes the baseline solutions to the four tasks of the Trauma THOMPSON Challenge 2023. They were not provided to the participants ahead of the challenge and was not part of the competition. The Temporal Segmentation Network algorithm is used for the action recognition task with top-1 accuracy of 11.38%. It is also used for the action anticipation task with top-1 accuracy of 5.82% and procedure recognition task with top-1 accuracy of 53.57%. ViLT is adopted for the Visual Question Answering challenge with an aggregate accuracy of 39.13%.

KW - Action Anticipation

KW - Action Recognition

KW - Procedure Recognition

KW - Visual Question Answering

UR - http://www.scopus.com/inward/record.url?scp=85208436615&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-71626-3_8

DO - 10.1007/978-3-031-71626-3_8

M3 - Conference contribution

AN - SCOPUS:85208436615

SN - 9783031716256

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

SP - 61

EP - 71

BT - AI for Brain Lesion Detection and Trauma Video Action Recognition - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge, Held in Conjunction with MICCAI 2023, Proceedings

A2 - Bao, Rina

A2 - Grant, Ellen

A2 - Ou, Yangming

A2 - Kirkpatrick, Andrew

A2 - Wachs, Juan

PB - Springer Science and Business Media Deutschland GmbH

T2 - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge Held in Conjunction with 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023

Y2 - 12 October 2023 through 16 October 2023

ER -

Zhuo Y, W. Kirkpatrick A, Couperus K, Tran O, Wachs J. The Trauma THOMPSON Challenge Report MICCAI 2023. In Bao R, Grant E, Ou Y, Kirkpatrick A, Wachs J, editors, AI for Brain Lesion Detection and Trauma Video Action Recognition - 1st BONBID-HIE Lesion Segmentation Challenge and 1st Trauma Thompson Challenge, Held in Conjunction with MICCAI 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2025. p. 61-71. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-71626-3_8

The Trauma THOMPSON Challenge Report MICCAI 2023

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this