Beyond rankings: learning (more) from algorithm validation

Challenges have become the state-of-the-art approach to benchmark image analysis algorithms in a comparative manner. While the validation on identical data sets was a great step forward, results analysis is often restricted to pure ranking tables, leaving relevant questions unanswered. Specifically,...

Full description

Saved in:
Bibliographic Details
Main Authors: Roß, Tobias (Author) , Bruno, Pierangela (Author) , Reinke, Annika (Author) , Wiesenfarth, Manuel (Author) , Köppel, Lisa (Author) , Full, Peter M. (Author) , Pekdemir, Bünyamin (Author) , Godau, Patrick (Author) , Trofimova, Darya (Author) , Isensee, Fabian (Author) , Adler, Tim (Author) , Tran, Thuy (Author) , Moccia, Sara (Author) , Calimeri, Francesco (Author) , Müller, Beat P. (Author) , Kopp-Schneider, Annette (Author) , Maier-Hein, Lena (Author)
Format: Article (Journal)
Language:English
Published: 23 March 2023
In: Medical image analysis
Year: 2023, Volume: 86, Pages: 1-12
ISSN:1361-8423
DOI:10.1016/j.media.2023.102765
Online Access:Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1016/j.media.2023.102765
Verlag, lizenzpflichtig, Volltext: https://www.sciencedirect.com/science/article/pii/S1361841523000269
Get full text
Author Notes:Tobias Roß, Pierangela Bruno, Annika Reinke, Manuel Wiesenfarth, Lisa Koeppel, Peter M. Full, Bünyamin Pekdemir, Patrick Godau, Darya Trofimova, Fabian Isensee, Tim J. Adler, Thuy N. Tran, Sara Moccia, Francesco Calimeri, Beat P. Müller-Stich, Annette Kopp-Schneider, Lena Maier-Hein

MARC

LEADER 00000caa a2200000 c 4500
001 1850719098
003 DE-627
005 20250901171110.0
007 cr uuu---uuuuu
008 230621s2023 xx |||||o 00| ||eng c
024 7 |a 10.1016/j.media.2023.102765  |2 doi 
035 |a (DE-627)1850719098 
035 |a (DE-599)KXP1850719098 
035 |a (OCoLC)1389528154 
040 |a DE-627  |b ger  |c DE-627  |e rda 
041 |a eng 
084 |a 33  |2 sdnb 
100 1 |a Roß, Tobias  |d 1990-  |e VerfasserIn  |0 (DE-588)118364955X  |0 (DE-627)1663271143  |4 aut 
245 1 0 |a Beyond rankings  |b learning (more) from algorithm validation  |c Tobias Roß, Pierangela Bruno, Annika Reinke, Manuel Wiesenfarth, Lisa Koeppel, Peter M. Full, Bünyamin Pekdemir, Patrick Godau, Darya Trofimova, Fabian Isensee, Tim J. Adler, Thuy N. Tran, Sara Moccia, Francesco Calimeri, Beat P. Müller-Stich, Annette Kopp-Schneider, Lena Maier-Hein 
264 1 |c 23 March 2023 
300 |a 12 
336 |a Text  |b txt  |2 rdacontent 
337 |a Computermedien  |b c  |2 rdamedia 
338 |a Online-Ressource  |b cr  |2 rdacarrier 
500 |a Online verfügbar 1. März 2023, Artikelversion 23. März 2023 
500 |a Gesehen am 21.06.2023 
520 |a Challenges have become the state-of-the-art approach to benchmark image analysis algorithms in a comparative manner. While the validation on identical data sets was a great step forward, results analysis is often restricted to pure ranking tables, leaving relevant questions unanswered. Specifically, little effort has been put into the systematic investigation on what characterizes images in which state-of-the-art algorithms fail. To address this gap in the literature, we (1) present a statistical framework for learning from challenges and (2) instantiate it for the specific task of instrument instance segmentation in laparoscopic videos. Our framework relies on the semantic meta data annotation of images, which serves as foundation for a General Linear Mixed Models (GLMM) analysis. Based on 51,542 meta data annotations performed on 2,728 images, we applied our approach to the results of the Robust Medical Instrument Segmentation Challenge (ROBUST-MIS) challenge 2019 and revealed underexposure, motion and occlusion of instruments as well as the presence of smoke or other objects in the background as major sources of algorithm failure. Our subsequent method development, tailored to the specific remaining issues, yielded a deep learning model with state-of-the-art overall performance and specific strengths in the processing of images in which previous methods tended to fail. Due to the objectivity and generic applicability of our approach, it could become a valuable tool for validation in the field of medical image analysis and beyond. 
650 4 |a Artificial intelligence 
650 4 |a Biomedical image analysis challenges 
650 4 |a Deep learning 
650 4 |a Endoscopic vision 
650 4 |a Generalized linear mixed models 
650 4 |a Grand challenges 
650 4 |a Image characteristics driven algorithm development 
650 4 |a Instrument segmentation 
650 4 |a Minimally invasive surgery 
650 4 |a Surgical data science 
700 1 |a Bruno, Pierangela  |e VerfasserIn  |4 aut 
700 1 |a Reinke, Annika  |d 1993-  |e VerfasserIn  |0 (DE-588)1219684252  |0 (DE-627)1735662143  |4 aut 
700 1 |a Wiesenfarth, Manuel  |e VerfasserIn  |4 aut 
700 1 |a Köppel, Lisa  |e VerfasserIn  |0 (DE-588)1222994089  |0 (DE-627)1742265537  |4 aut 
700 1 |a Full, Peter M.  |e VerfasserIn  |0 (DE-588)1219695777  |0 (DE-627)1735698784  |4 aut 
700 1 |a Pekdemir, Bünyamin  |e VerfasserIn  |4 aut 
700 1 |a Godau, Patrick  |d 1993-  |e VerfasserIn  |0 (DE-588)1179282515  |0 (DE-627)1066788200  |0 (DE-576)518174212  |4 aut 
700 1 |a Trofimova, Darya  |e VerfasserIn  |0 (DE-588)1062761766  |0 (DE-627)806884908  |0 (DE-576)420194851  |4 aut 
700 1 |a Isensee, Fabian  |d 1990-  |e VerfasserIn  |0 (DE-588)1207568430  |0 (DE-627)1694044998  |4 aut 
700 1 |a Adler, Tim  |d 1991-  |e VerfasserIn  |0 (DE-588)1194987672  |0 (DE-627)1677181680  |4 aut 
700 1 |a Tran, Thuy  |e VerfasserIn  |0 (DE-588)1293398039  |0 (DE-627)1850720029  |4 aut 
700 1 |a Moccia, Sara  |e VerfasserIn  |4 aut 
700 1 |a Calimeri, Francesco  |e VerfasserIn  |4 aut 
700 1 |a Müller, Beat P.  |d 1971-  |e VerfasserIn  |0 (DE-588)14066209X  |0 (DE-627)70374819X  |0 (DE-576)317992287  |4 aut 
700 1 |a Kopp-Schneider, Annette  |e VerfasserIn  |0 (DE-588)1119160545  |0 (DE-627)872460444  |0 (DE-576)178153206  |4 aut 
700 1 |a Maier-Hein, Lena  |d 1980-  |e VerfasserIn  |0 (DE-588)1075029252  |0 (DE-627)832869899  |0 (DE-576)190090804  |4 aut 
773 0 8 |i Enthalten in  |t Medical image analysis  |d Amsterdam [u.a.] : Elsevier Science, 1996  |g 86(2023) vom: März, Artikel-ID 102765, Seite 1-12  |h Online-Ressource  |w (DE-627)306365081  |w (DE-600)1497450-2  |w (DE-576)091204941  |x 1361-8423  |7 nnas  |a Beyond rankings learning (more) from algorithm validation 
773 1 8 |g volume:86  |g year:2023  |g month:03  |g elocationid:102765  |g pages:1-12  |g extent:12  |a Beyond rankings learning (more) from algorithm validation 
856 4 0 |u https://doi.org/10.1016/j.media.2023.102765  |x Verlag  |x Resolving-System  |z lizenzpflichtig  |3 Volltext 
856 4 0 |u https://www.sciencedirect.com/science/article/pii/S1361841523000269  |x Verlag  |z lizenzpflichtig  |3 Volltext 
951 |a AR 
992 |a 20230621 
993 |a Article 
994 |a 2023 
998 |g 1075029252  |a Maier-Hein, Lena  |m 1075029252:Maier-Hein, Lena  |d 110000  |e 110000PM1075029252  |k 0/110000/  |p 17  |y j 
998 |g 1119160545  |a Kopp-Schneider, Annette  |m 1119160545:Kopp-Schneider, Annette  |d 50000  |e 50000PK1119160545  |k 0/50000/  |p 16 
998 |g 14066209X  |a Müller, Beat P.  |m 14066209X:Müller, Beat P.  |d 50000  |e 50000PM14066209X  |k 0/50000/  |p 15 
998 |g 1293398039  |a Tran, Thuy  |m 1293398039:Tran, Thuy  |d 110000  |e 110000PT1293398039  |k 0/110000/  |p 12 
998 |g 1194987672  |a Adler, Tim  |m 1194987672:Adler, Tim  |p 11 
998 |g 1207568430  |a Isensee, Fabian  |m 1207568430:Isensee, Fabian  |p 10 
998 |g 1062761766  |a Trofimova, Darya  |m 1062761766:Trofimova, Darya  |p 9 
998 |g 1179282515  |a Godau, Patrick  |m 1179282515:Godau, Patrick  |d 110000  |e 110000PG1179282515  |k 0/110000/  |p 8 
998 |g 1219695777  |a Full, Peter M.  |m 1219695777:Full, Peter M.  |d 50000  |e 50000PF1219695777  |k 0/50000/  |p 6 
998 |g 1222994089  |a Köppel, Lisa  |m 1222994089:Köppel, Lisa  |d 910000  |d 911700  |e 910000PK1222994089  |e 911700PK1222994089  |k 0/910000/  |k 1/910000/911700/  |p 5 
998 |g 1219684252  |a Reinke, Annika  |m 1219684252:Reinke, Annika  |d 50000  |e 50000PR1219684252  |k 0/50000/  |p 3 
999 |a KXP-PPN1850719098  |e 4341189123 
BIB |a Y 
SER |a journal 
JSO |a {"recId":"1850719098","id":{"doi":["10.1016/j.media.2023.102765"],"eki":["1850719098"]},"title":[{"subtitle":"learning (more) from algorithm validation","title":"Beyond rankings","title_sort":"Beyond rankings"}],"note":["Online verfügbar 1. März 2023, Artikelversion 23. März 2023","Gesehen am 21.06.2023"],"person":[{"family":"Roß","roleDisplay":"VerfasserIn","display":"Roß, Tobias","role":"aut","given":"Tobias"},{"family":"Bruno","roleDisplay":"VerfasserIn","given":"Pierangela","role":"aut","display":"Bruno, Pierangela"},{"display":"Reinke, Annika","role":"aut","given":"Annika","family":"Reinke","roleDisplay":"VerfasserIn"},{"roleDisplay":"VerfasserIn","family":"Wiesenfarth","role":"aut","display":"Wiesenfarth, Manuel","given":"Manuel"},{"family":"Köppel","roleDisplay":"VerfasserIn","role":"aut","display":"Köppel, Lisa","given":"Lisa"},{"role":"aut","display":"Full, Peter M.","given":"Peter M.","roleDisplay":"VerfasserIn","family":"Full"},{"roleDisplay":"VerfasserIn","family":"Pekdemir","display":"Pekdemir, Bünyamin","role":"aut","given":"Bünyamin"},{"display":"Godau, Patrick","role":"aut","given":"Patrick","roleDisplay":"VerfasserIn","family":"Godau"},{"given":"Darya","role":"aut","display":"Trofimova, Darya","roleDisplay":"VerfasserIn","family":"Trofimova"},{"roleDisplay":"VerfasserIn","family":"Isensee","given":"Fabian","role":"aut","display":"Isensee, Fabian"},{"role":"aut","display":"Adler, Tim","given":"Tim","family":"Adler","roleDisplay":"VerfasserIn"},{"roleDisplay":"VerfasserIn","family":"Tran","given":"Thuy","display":"Tran, Thuy","role":"aut"},{"family":"Moccia","roleDisplay":"VerfasserIn","given":"Sara","role":"aut","display":"Moccia, Sara"},{"given":"Francesco","display":"Calimeri, Francesco","role":"aut","family":"Calimeri","roleDisplay":"VerfasserIn"},{"display":"Müller, Beat P.","role":"aut","given":"Beat P.","family":"Müller","roleDisplay":"VerfasserIn"},{"family":"Kopp-Schneider","roleDisplay":"VerfasserIn","given":"Annette","role":"aut","display":"Kopp-Schneider, Annette"},{"given":"Lena","display":"Maier-Hein, Lena","role":"aut","family":"Maier-Hein","roleDisplay":"VerfasserIn"}],"name":{"displayForm":["Tobias Roß, Pierangela Bruno, Annika Reinke, Manuel Wiesenfarth, Lisa Koeppel, Peter M. Full, Bünyamin Pekdemir, Patrick Godau, Darya Trofimova, Fabian Isensee, Tim J. Adler, Thuy N. Tran, Sara Moccia, Francesco Calimeri, Beat P. Müller-Stich, Annette Kopp-Schneider, Lena Maier-Hein"]},"physDesc":[{"extent":"12 S."}],"relHost":[{"recId":"306365081","titleAlt":[{"title":"Medical image analysis online"}],"pubHistory":["1.1996/97 -"],"language":["eng"],"origin":[{"publisher":"Elsevier Science","dateIssuedKey":"1996","publisherPlace":"Amsterdam [u.a.]","dateIssuedDisp":"1996-"}],"type":{"bibl":"periodical","media":"Online-Ressource"},"part":{"year":"2023","extent":"12","pages":"1-12","text":"86(2023) vom: März, Artikel-ID 102765, Seite 1-12","volume":"86"},"physDesc":[{"extent":"Online-Ressource"}],"title":[{"title":"Medical image analysis","title_sort":"Medical image analysis"}],"id":{"issn":["1361-8423"],"zdb":["1497450-2"],"eki":["306365081"]},"disp":"Beyond rankings learning (more) from algorithm validationMedical image analysis","note":["Gesehen am 16.05.23"]}],"type":{"bibl":"article-journal","media":"Online-Ressource"},"origin":[{"dateIssuedKey":"2023","dateIssuedDisp":"23 March 2023"}],"language":["eng"]} 
SRT |a ROSSTOBIASBEYONDRANK2320