Papers › PEg TRAnsfer Workflow recognition challenge report: Does multi-modal data improve recognition?

PEg TRAnsfer Workflow recognition challenge report: Does multi-modal data improve recognition?

11 Feb 2022arXiv:2202.05821archive 2025-07-28

Arnaud Huaulmé, Kanako Harada, Quang-Minh Nguyen, Bogyu Park, Seungbum Hong, Min-Kook Choi, Michael Peven, Yunshuang Li, Yonghao Long, Qi Dou, Satyadwyoom Kumar, Seenivasan Lalithkumar, Ren Hongliang, Hiroki Matsuzaki, Yuto Ishikawa, Yuriko Harai, Satoshi Kondo, Mamoru Mitsuishi, Pierre Jannin

This paper presents the design and results of the "PEg TRAnsfert Workflow recognition" (PETRAW) challenge whose objective was to develop surgical workflow recognition methods based on one or several modalities, among video, kinematic, and segmentation data, in order to study their added value. The PETRAW challenge provided a data set of 150 peg transfer sequences performed on a virtual simulator. This data set was composed of videos, kinematics, semantic segmentation, and workflow annotations which described the sequences at three different granularity levels: phase, step, and activity. Five tasks were proposed to the participants: three of them were related to the recognition of all granularities with one of the available modalities, while the others addressed the recognition with a combination of modalities. Average application-dependent balanced accuracy (AD-Accuracy) was used as evaluation metric to take unbalanced classes into account and because it is more clinically relevant than a frame-by-frame score. Seven teams participated in at least one task and four of them in all tasks. Best results are obtained with the use of the video and the kinematics data with an AD-Accuracy between 93% and 90% for the four teams who participated in all tasks. The improvement between video/kinematic-based methods and the uni-modality ones was significant for all of the teams. However, the difference in testing execution time between the video/kinematic-based and the kinematic-based methods has to be taken into consideration. Is it relevant to spend 20 to 200 times more computing time for less than 3% of improvement? The PETRAW data set is publicly available at www.synapse.org/PETRAW to encourage further research in surgical workflow recognition.

PaperPDF

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Kinematic Based Workflow RecognitionSegmentation Based Workflow RecognitionSemantic SegmentationVideo & Kinematic Base Workflow RecognitionVideo Based Workflow RecognitionVideo, Kinematic & Segmentation Base Workflow Recognition

Datasets

Introduced by this paper, per the archive.

PETRAW

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Kinematic Based Workflow Recognition PETRAW MedAIR Average AD-Accuracy 90.72 #1 of 6 Archive leaderboard report
Kinematic Based Workflow Recognition PETRAW NCC Next Average AD-Accuracy 90.32 #2 of 6 Archive leaderboard report
Kinematic Based Workflow Recognition PETRAW MediCIS Average AD-Accuracy 89.71 #3 of 6 Archive leaderboard report
Kinematic Based Workflow Recognition PETRAW SK Average AD-Accuracy 89.66 #4 of 6 Archive leaderboard report
Kinematic Based Workflow Recognition PETRAW JHU-CIRL Average AD-Accuracy 86.45 #5 of 6 Archive leaderboard report
Kinematic Based Workflow Recognition PETRAW Hutom Average AD-Accuracy 84.31 #6 of 6 Archive leaderboard report
Segmentation Based Workflow Recognition PETRAW SK Average AD-Accuracy 88.51 #1 of 4 Archive leaderboard report
Segmentation Based Workflow Recognition PETRAW NCC Next Average AD-Accuracy 87.71 #2 of 4 Archive leaderboard report
Segmentation Based Workflow Recognition PETRAW MediCIS Average AD-Accuracy 87.22 #3 of 4 Archive leaderboard report
Segmentation Based Workflow Recognition PETRAW Hutom Average AD-Accuracy 60.28 #4 of 4 Archive leaderboard report
Semantic Segmentation PETRAW NCC Next Mean IoU (class) 96.9 #1 of 4 Archive leaderboard report
Semantic Segmentation PETRAW SK Mean IoU (class) 96.4 #2 of 4 Archive leaderboard report
Semantic Segmentation PETRAW MediCIS Mean IoU (class) 94 #3 of 4 Archive leaderboard report
Semantic Segmentation PETRAW Hutom Mean IoU (class) 85 #4 of 4 Archive leaderboard report
Video & Kinematic Base Workflow Recognition PETRAW NCC Next Average AD-Accuracy 93.09 #1 of 6 Archive leaderboard report
Video & Kinematic Base Workflow Recognition PETRAW SK Average AD-Accuracy 91.61 #2 of 6 Archive leaderboard report
Video & Kinematic Base Workflow Recognition PETRAW Hutom Average AD-Accuracy 91.33 #3 of 6 Archive leaderboard report
Video & Kinematic Base Workflow Recognition PETRAW MediCIS Average AD-Accuracy 90.18 #4 of 6 Archive leaderboard report
Video & Kinematic Base Workflow Recognition PETRAW MedAIR Average AD-Accuracy 86.98 #5 of 6 Archive leaderboard report
Video & Kinematic Base Workflow Recognition PETRAW MMLAB Average AD-Accuracy 84.8 #6 of 6 Archive leaderboard report
Video Based Workflow Recognition PETRAW SK Average AD-Accuracy 90.77 #1 of 5 Archive leaderboard report
Video Based Workflow Recognition PETRAW Hutom Average AD-Accuracy 90.51 #2 of 5 Archive leaderboard report
Video Based Workflow Recognition PETRAW MediCIS Average AD-Accuracy 89.15 #3 of 5 Archive leaderboard report
Video Based Workflow Recognition PETRAW NCC Next Average AD-Accuracy 87.77 #4 of 5 Archive leaderboard report
Video Based Workflow Recognition PETRAW MedAIR Average AD-Accuracy 84.31 #5 of 5 Archive leaderboard report
Video, Kinematic & Segmentation Base Workflow Recognition PETRAW NCC Next Average AD-Accuracy 93.09 #1 of 4 Archive leaderboard report
Video, Kinematic & Segmentation Base Workflow Recognition PETRAW SK Average AD-Accuracy 91.37 #2 of 4 Archive leaderboard report
Video, Kinematic & Segmentation Base Workflow Recognition PETRAW Hutom Average AD-Accuracy 91.27 #3 of 4 Archive leaderboard report
Video, Kinematic & Segmentation Base Workflow Recognition PETRAW MediCIS Task 5 Average AD-Accuracy 89.81 #4 of 4 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections