Comparative validation of machine learning algorithms for surgical workflow and skill analysis with the HeiChole benchmark

Aug 5, 2026

Upload by Finn

Surgical workflow and skill analysis underpin next-generation cognitive surgical assistance systems, enabling context-sensitive warnings, semi-autonomous robotic support, and data-driven training feedback. Building on prior phase-recognition results on open single-center video data, this study examines whether recognition algorithms generalize across multiple centers. The investigation targets more challenging tasks beyond phases, including surgical action identification and surgical skill assessment. Evaluation is conducted using the HeiChole benchmark to compare algorithm performance under multicenter conditions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages21

Reading Time53 min

AccessFree to download

Abstract

Purpose

Methods

Evaluation

Medical Image Analysis 86 (2023) 102770 Contents lists available at ScienceDirect Medical Image Analysis journal [homepage:](homepage: www.elsevier.com/locate/media)[ www.elsevier.com/locate/media](homepage: www.elsevier.com/locate/media) Comparative validation of machine learning algorithms for surgical workflow and skill analysis with the HeiChole benchmark Martin Wagner a, b, *, Beat-Peter Müller-Stich a, b, Anna Kisilenko a, b, Duc Tran a, b, Patrick Heger a, Lars Mündermann c, David M Lubotsky a, b, Benjamin Müller a, b, Tornike Davitashvili a, b, Manuela Capek a, b, Annika Reinked, e, f, Carissa Reid g, Tong Yu h, i, Armine Vardazaryan h, i, Chinedu Innocent Nwoyeh, i, Nicolas Padoy h, i, Xinyang Liuj, Eung-Joo Leek, Constantin Dischl, Hans Meinel, m, Tong Xia n, Fucang Jia n, Satoshi Kondo o, 2, Wolfgang Reiter p, Yueming Jin q, Yonghao Long q, Meirui Jiang q, Qi Dou q, Pheng Ann Heng q, Isabell Twick r, Kadir Kirtac r, Enes Hosgor r, Jon Lindstr¨om Bolmgren r, Michael Stenzel r, Bj¨orn von Siemens r, Long Zhao s, Zhenxiao Ge s, Haiming Sun s, Di Xie s, Mengqi Guot, Daochang Liu u, Hannes G. Kenngott a, Felix Nickel a, Moritz von Frankenberg v, Franziska Mathis-Ullrich w, Annette Kopp-Schneider g, Lena Maier-Hein d, e, f, x, Stefanie Speidel y, z, 1, Sebastian Bodenstedty, z, 1 a Department for General, Visceral and Transplantation Surgery, Heidelberg University Hospital, Im Neuenheimer Feld 420, 69120 Heidelberg, Germany b National Center for Tumor Diseases (NCT) Heidelberg, Im Neuenheimer Feld 460, 69120 Heidelberg, Germany c Data Assisted Solutions, Corporate Research & Technology, KARL STORZ SE & Co. KG, Dr. Karl-Storz-Str. 34, 78332 Tuttlingen d Div. Computer Assisted Medical Interventions, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 223, 69120 Heidelberg Germany e HIP Helmholtz Imaging Platform, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 223, 69120 Heidelberg Germany f Faculty of Mathematics and Computer Science, Heidelberg University, Im Neuenheimer Feld 205, 69120 Heidelberg g Division of Biostatistics, German Cancer Research Center, Im Neuenheimer Feld 280, Heidelberg, Germany h ICube, University of Strasbourg, CNRS, France. 300 bd S´ebastien Brant-CS 10413, F-67412 Illkirch Cedex, France i IHU Strasbourg, France. 1 Place de l’hˆopital, 67000 Strasbourg, France j Sheikh Zayed Institute for Pediatric Surgical Innovation, Children’s National Hospital, 111 Michigan Ave NW, Washington, DC 20010, USA k University of Maryland, College Park, 2405 A V Williams Building, College Park, MD 20742, USA l Fraunhofer Institute for Digital Medicine MEVIS, Max-von-Laue-Str. 2, 28359 Bremen, Germany m University of Bremen, FB3, Medical Image Computing Group, ℅ Fraunhofer MEVIS, Am Fallturm 1, 28359 Bremen, Germany n Lab for Medical Imaging and Digital Surgery, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, China o Konika Minolta, Inc., 1-2, Sakura-machi, Takatsuki, Oasak 569-8503, Japan p Wintegral GmbH, Ehrenbreitsteiner Str. 36, 80993 München, Germany q Department of Computer Science and Engineering, Ho Sin-Hang Engineering Building, The Chinese University of Hong Kong, Sha Tin, NT, Hong Kong r Caresyntax GmbH, Komturstr. 18A, 12099 Berlin, Germany s Hikvision Research Institute, Hangzhou, China t School of Computing, National University of Singapore, Computing 1, No.13 Computing Drive, 117417, Singapore u National Engineering Research Center of Visual Technology, School of Computer Science, Peking University, Beijing, China v Department of Surgery, Salem Hospital of the Evangelische Stadtmission Heidelberg, Zeppelinstrasse 11-33, 69121 Heidelberg, Germany w Health Robotics and Automation Laboratory, Institute for Anthropomatics and Robotics, Karlsruhe Institute of Technology, Geb. 40.28, KIT Campus Süd, Engler-BunteRing 8, 76131 Karlsruhe, Germany x Medical Faculty, Heidelberg University, Im Neuenheimer Feld 672, 69120

FAQs

What is the main goal of this study?

To assess how well surgical workflow and skill recognition algorithms generalize when moving from single-center data to a multicenter setting using the HeiChole benchmark.

Which tasks are evaluated beyond phase recognition?

The study evaluates more difficult recognition tasks, including surgical action recognition and surgical skill assessment.

Why is surgical workflow and skill analysis important?

It enables cognitive surgical assistance such as context-sensitive warnings, semi-autonomous robotic help, and data-driven feedback for training surgeons.

Aug 5, 2026

Upload by Finn