Comparison of image normalization methods for multi-site deep learning

In this study, we evaluate the influence of normalization on the performance of deep learning networks for tumor segmentation and the prediction of the pathological response of locally advanced rectal cancer to neoadjuvant chemoradiotherapy. The techniques were applied to a multicenter and multimoda...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Albert, Steffen (VerfasserIn) , Wichtmann, Barbara (VerfasserIn) , Zhao, Wenzhao (VerfasserIn) , Maurer, Angelika (VerfasserIn) , Hesser, Jürgen (VerfasserIn) , Attenberger, Ulrike (VerfasserIn) , Schad, Lothar R. (VerfasserIn) , Zöllner, Frank G. (VerfasserIn)
Dokumenttyp: Article (Journal)
Sprache:Englisch
Veröffentlicht: 2023
In: Applied Sciences
Year: 2023, Jahrgang: 13, Heft: 15, Pages: 1-13
ISSN:2076-3417
DOI:10.3390/app13158923
Online-Zugang:Verlag, kostenfrei, Volltext: https://doi.org/10.3390/app13158923
Verlag, kostenfrei, Volltext: https://www.mdpi.com/2076-3417/13/15/8923
Volltext
Verfasserangaben:Steffen Albert, Barbara D. Wichtmann, Wenzhao Zhao, Angelika Maurer, Jürgen Hesser, Ulrike I. Attenberger, Lothar R. Schad and Frank G. Zöllner

MARC

LEADER 00000caa a2200000 c 4500
001 1860346987
003 DE-627
005 20240307052841.0
007 cr uuu---uuuuu
008 230927s2023 xx |||||o 00| ||eng c
024 7 |a 10.3390/app13158923  |2 doi 
035 |a (DE-627)1860346987 
035 |a (DE-599)KXP1860346987 
035 |a (OCoLC)1425210945 
040 |a DE-627  |b ger  |c DE-627  |e rda 
041 |a eng 
084 |a 33  |2 sdnb 
100 1 |a Albert, Steffen  |d 1993-  |e VerfasserIn  |0 (DE-588)1211680568  |0 (DE-627)1700313290  |4 aut 
245 1 0 |a Comparison of image normalization methods for multi-site deep learning  |c Steffen Albert, Barbara D. Wichtmann, Wenzhao Zhao, Angelika Maurer, Jürgen Hesser, Ulrike I. Attenberger, Lothar R. Schad and Frank G. Zöllner 
264 1 |c 2023 
300 |a 13 
336 |a Text  |b txt  |2 rdacontent 
337 |a Computermedien  |b c  |2 rdamedia 
338 |a Online-Ressource  |b cr  |2 rdacarrier 
500 |a Veröffentlicht: 3. August 2023 
500 |a Gesehen am 27.09.2023 
520 |a In this study, we evaluate the influence of normalization on the performance of deep learning networks for tumor segmentation and the prediction of the pathological response of locally advanced rectal cancer to neoadjuvant chemoradiotherapy. The techniques were applied to a multicenter and multimodal magnet resonance imaging data set consisting of 201 patients recorded at six centers. We implemented and investigated six different normalization methods (setting the mean and standard deviation, histogram matching, percentiles, combining percentiles and histogram matching, fixed window and an auto-encoder with adversarial loss using the imaging parameters) and evaluated their impact on four deep learning tasks: tumor segmentation, prediction of treatment outcome, and prediction of sex and age. The latter two tasks were implemented as a reference test. We trained a modified U-Net with different normalization methods in multiple configurations: on all images, images from all centers except one, and images from a single center. Our results show that normalization only plays a minor role in segmentation, with a difference in Dice of less than 0.02 between the best and worst performing networks. For the prediction of sex and treatment outcomes, the percentile method combined with histogram matching works best for all scenarios. The biggest difference in performance, depending on the normalization method, occurs for classification. In conclusion, normalization is especially important for small data sets or for generalizing to different data distributions. The deep learning method was superior to the classical methods only in a minority of cases, probably due to the limited amount of training data. 
650 4 |a medical imaging 
650 4 |a MRI 
650 4 |a normalization 
700 1 |a Wichtmann, Barbara  |d 1992-  |e VerfasserIn  |0 (DE-588)113837461X  |0 (DE-627)895889684  |0 (DE-576)492514135  |4 aut 
700 1 |a Zhao, Wenzhao  |d 1990-  |e VerfasserIn  |0 (DE-588)1304109860  |0 (DE-627)186034769X  |4 aut 
700 1 |a Maurer, Angelika  |e VerfasserIn  |4 aut 
700 1 |a Hesser, Jürgen  |d 1964-  |e VerfasserIn  |0 (DE-588)1020647353  |0 (DE-627)691291071  |0 (DE-576)361513739  |4 aut 
700 1 |a Attenberger, Ulrike  |d 1980-  |e VerfasserIn  |0 (DE-588)131554859  |0 (DE-627)510698557  |0 (DE-576)298582775  |4 aut 
700 1 |a Schad, Lothar R.  |d 1956-  |e VerfasserIn  |0 (DE-588)1028817630  |0 (DE-627)731640241  |0 (DE-576)376271221  |4 aut 
700 1 |a Zöllner, Frank G.  |d 1976-  |e VerfasserIn  |0 (DE-588)129580015  |0 (DE-627)473357054  |0 (DE-576)297732587  |4 aut 
773 0 8 |i Enthalten in  |t Applied Sciences  |d Basel : MDPI, 2011  |g 13(2023), 15, Artikel-ID 8923, Seite 1-13  |h Online-Ressource  |w (DE-627)737287640  |w (DE-600)2704225-X  |w (DE-576)379466716  |x 2076-3417  |7 nnas  |a Comparison of image normalization methods for multi-site deep learning 
773 1 8 |g volume:13  |g year:2023  |g number:15  |g elocationid:8923  |g pages:1-13  |g extent:13  |a Comparison of image normalization methods for multi-site deep learning 
856 4 0 |u https://doi.org/10.3390/app13158923  |x Verlag  |x Resolving-System  |z kostenfrei  |3 Volltext 
856 4 0 |u https://www.mdpi.com/2076-3417/13/15/8923  |x Verlag  |z kostenfrei  |3 Volltext 
951 |a AR 
992 |a 20230927 
993 |a Article 
994 |a 2023 
998 |g 129580015  |a Zöllner, Frank G.  |m 129580015:Zöllner, Frank G.  |d 60000  |d 60000  |d 65200  |e 60000PZ129580015  |e 60000PZ129580015  |e 65200PZ129580015  |k 0/60000/  |k 0/60000/  |k 1/60000/65200/  |p 8  |y j 
998 |g 1028817630  |a Schad, Lothar R.  |m 1028817630:Schad, Lothar R.  |d 60000  |d 65200  |e 60000PS1028817630  |e 65200PS1028817630  |k 0/60000/  |k 1/60000/65200/  |p 7 
998 |g 131554859  |a Attenberger, Ulrike  |m 131554859:Attenberger, Ulrike  |d 60000  |e 60000PA131554859  |k 0/60000/  |p 6 
998 |g 1020647353  |a Hesser, Jürgen  |m 1020647353:Hesser, Jürgen  |d 60000  |d 65200  |e 60000PH1020647353  |e 65200PH1020647353  |k 0/60000/  |k 1/60000/65200/  |p 5 
998 |g 1304109860  |a Zhao, Wenzhao  |m 1304109860:Zhao, Wenzhao  |d 60000  |d 63000  |e 60000PZ1304109860  |e 63000PZ1304109860  |k 0/60000/  |k 1/60000/63000/  |p 3 
998 |g 113837461X  |a Wichtmann, Barbara  |m 113837461X:Wichtmann, Barbara  |d 60000  |e 60000PW113837461X  |k 0/60000/  |p 2 
998 |g 1211680568  |a Albert, Steffen  |m 1211680568:Albert, Steffen  |d 60000  |d 65200  |e 60000PA1211680568  |e 65200PA1211680568  |k 0/60000/  |k 1/60000/65200/  |p 1  |x j 
999 |a KXP-PPN1860346987  |e 4379470776 
BIB |a Y 
SER |a journal 
JSO |a {"type":{"bibl":"article-journal","media":"Online-Ressource"},"origin":[{"dateIssuedDisp":"2023","dateIssuedKey":"2023"}],"note":["Veröffentlicht: 3. August 2023","Gesehen am 27.09.2023"],"title":[{"title":"Comparison of image normalization methods for multi-site deep learning","title_sort":"Comparison of image normalization methods for multi-site deep learning"}],"id":{"doi":["10.3390/app13158923"],"eki":["1860346987"]},"name":{"displayForm":["Steffen Albert, Barbara D. Wichtmann, Wenzhao Zhao, Angelika Maurer, Jürgen Hesser, Ulrike I. Attenberger, Lothar R. Schad and Frank G. Zöllner"]},"physDesc":[{"extent":"13 S."}],"relHost":[{"part":{"extent":"13","volume":"13","issue":"15","text":"13(2023), 15, Artikel-ID 8923, Seite 1-13","year":"2023","pages":"1-13"},"language":["eng"],"id":{"issn":["2076-3417"],"zdb":["2704225-X"],"eki":["737287640"]},"physDesc":[{"extent":"Online-Ressource"}],"recId":"737287640","disp":"Comparison of image normalization methods for multi-site deep learningApplied Sciences","type":{"media":"Online-Ressource","bibl":"periodical"},"title":[{"title":"Applied Sciences","subtitle":"open access journal","title_sort":"Applied Sciences"}],"origin":[{"dateIssuedKey":"2011","publisher":"MDPI","dateIssuedDisp":"2011-","publisherPlace":"Basel"}],"note":["Gesehen am 19.02.13"],"pubHistory":["1.2011 -"]}],"recId":"1860346987","person":[{"given":"Steffen","role":"aut","display":"Albert, Steffen","family":"Albert"},{"role":"aut","given":"Barbara","display":"Wichtmann, Barbara","family":"Wichtmann"},{"display":"Zhao, Wenzhao","family":"Zhao","given":"Wenzhao","role":"aut"},{"given":"Angelika","role":"aut","family":"Maurer","display":"Maurer, Angelika"},{"display":"Hesser, Jürgen","family":"Hesser","given":"Jürgen","role":"aut"},{"family":"Attenberger","display":"Attenberger, Ulrike","role":"aut","given":"Ulrike"},{"family":"Schad","display":"Schad, Lothar R.","role":"aut","given":"Lothar R."},{"role":"aut","given":"Frank G.","family":"Zöllner","display":"Zöllner, Frank G."}],"language":["eng"]} 
SRT |a ALBERTSTEFCOMPARISON2023