Noise-augmented deep denoising: a method to boost CT image denoising networks

Background Denoising low dose computed tomography (CT) images can have great advantages for the aim of minimizing the radiation risk of the patients, as it can help lower the effective dose to the patient while providing constant image quality. In recent years, deep denoising methods became a popula...

Full description

Saved in:
Bibliographic Details
Main Authors: Kristof, Gernot (Author) , Eulig, Elias (Author) , Kachelrieß, Marc (Author)
Format: Article (Journal)
Language:English
Published: 23 September 2025
In: Medical physics
Year: 2025, Volume: 52, Issue: 10, Pages: 1-13
ISSN:2473-4209
DOI:10.1002/mp.18121
Online Access:Verlag, kostenfrei, Volltext: https://doi.org/10.1002/mp.18121
Verlag, kostenfrei, Volltext: https://onlinelibrary.wiley.com/doi/abs/10.1002/mp.18121
Get full text
Author Notes:Gernot Kristof, Elias Eulig, Marc Kachelrieß

MARC

LEADER 00000naa a2200000 c 4500
001 195019700X
003 DE-627
005 20260126132824.0
007 cr uuu---uuuuu
008 260126s2025 xx |||||o 00| ||eng c
024 7 |a 10.1002/mp.18121  |2 doi 
035 |a (DE-627)195019700X 
035 |a (DE-599)KXP195019700X 
040 |a DE-627  |b ger  |c DE-627  |e rda 
041 |a eng 
084 |a 33  |2 sdnb 
100 1 |a Kristof, Gernot  |e VerfasserIn  |0 (DE-588)1387987682  |0 (DE-627)1950210103  |4 aut 
245 1 0 |a Noise-augmented deep denoising  |b a method to boost CT image denoising networks  |c Gernot Kristof, Elias Eulig, Marc Kachelrieß 
264 1 |c 23 September 2025 
300 |b Illustration 
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 Gesehen am 26.01.2026 
520 |a Background Denoising low dose computed tomography (CT) images can have great advantages for the aim of minimizing the radiation risk of the patients, as it can help lower the effective dose to the patient while providing constant image quality. In recent years, deep denoising methods became a popular way to accomplish this task. Conventional deep denoising algorithms, however, cannot handle the correlation between neighboring pixels or voxels very well, because the noise structure in CT is a resultant of the global attenuation properties of the patient and because the receptive field of most denoising approaches is rather small. Purpose The purpose of this study is to improve existing denoising networks, by providing them additional information about the image noise. Methods We here propose to generate \N\ additional noise realizations by simulation, reconstruct them, and use these noise images as additional input into existing denoising networks. This noise augmentation is intended to guide the denoising process. The additional noise realizations are not only required during training, but also during inference. The rationale behind this noise-augmented deep denoising (NADD) is that CT image noise is strongly patient-specific and it is non-local since it depends on the attenuation of X-ray beams. NADD is architecture-agnostic and can thus be used to improve any previously proposed method. We demonstrate NADD using existing denoising networks that we slightly modified in their input layer in order to take the CT image that is to be denoised plus additional noise images as input. To do so, we modified three popular denoising networks, the CNN10, the ResNet, and the WGAN-VGG and apply them to clinical cases with 90% dose reduction. Results In all cases tested, the denoising networks strongly benefit from the noise augmentation. Noise artifacts that are being misinterpreted by the original networks as being anatomical structures, are correctly removed by the NADD version of the same networks. The more noise images are provided, the better the performance. Conclusions Providing additional simulated noise realizations helps to significantly improve the performance of CT image denoising networks. 
650 4 |a deep learning 
650 4 |a diagnostic CT 
650 4 |a dose reduction 
650 4 |a neural networks 
650 4 |a noise augmentation 
650 4 |a noise reduction 
700 1 |a Eulig, Elias  |d 1995-  |e VerfasserIn  |0 (DE-588)1191843181  |0 (DE-627)1670315185  |4 aut 
700 1 |a Kachelrieß, Marc  |d 1969-  |e VerfasserIn  |0 (DE-588)120866544  |0 (DE-627)705049280  |0 (DE-576)292422725  |4 aut 
773 0 8 |i Enthalten in  |t Medical physics  |d Hoboken, NJ : Wiley, 1974  |g 52(2025), 10 vom: Okt., Artikel-ID e18121, Seite 1-13  |h Online-Ressource  |w (DE-627)265784867  |w (DE-600)1466421-5  |w (DE-576)074891243  |x 2473-4209  |7 nnas  |a Noise-augmented deep denoising a method to boost CT image denoising networks 
773 1 8 |g volume:52  |g year:2025  |g number:10  |g month:10  |g elocationid:e18121  |g pages:1-13  |g extent:13  |a Noise-augmented deep denoising a method to boost CT image denoising networks 
856 4 0 |u https://doi.org/10.1002/mp.18121  |x Verlag  |x Resolving-System  |z kostenfrei  |3 Volltext  |7 0 
856 4 0 |u https://onlinelibrary.wiley.com/doi/abs/10.1002/mp.18121  |x Verlag  |z kostenfrei  |3 Volltext  |7 0 
951 |a AR 
992 |a 20260126 
993 |a Article 
994 |a 2025 
998 |g 120866544  |a Kachelrieß, Marc  |m 120866544:Kachelrieß, Marc  |d 50000  |e 50000PK120866544  |k 0/50000/  |p 3  |y j 
998 |g 1191843181  |a Eulig, Elias  |m 1191843181:Eulig, Elias  |p 2 
998 |g 1387987682  |a Kristof, Gernot  |m 1387987682:Kristof, Gernot  |p 1  |x j 
999 |a KXP-PPN195019700X  |e 4860735854 
BIB |a Y 
SER |a journal 
JSO |a {"person":[{"given":"Gernot","role":"aut","display":"Kristof, Gernot","family":"Kristof"},{"given":"Elias","role":"aut","display":"Eulig, Elias","family":"Eulig"},{"role":"aut","given":"Marc","display":"Kachelrieß, Marc","family":"Kachelrieß"}],"language":["eng"],"type":{"bibl":"article-journal","media":"Online-Ressource"},"note":["Gesehen am 26.01.2026"],"origin":[{"dateIssuedKey":"2025","dateIssuedDisp":"23 September 2025"}],"title":[{"subtitle":"a method to boost CT image denoising networks","title_sort":"Noise-augmented deep denoising","title":"Noise-augmented deep denoising"}],"recId":"195019700X","physDesc":[{"extent":"13 S.","noteIll":"Illustration"}],"name":{"displayForm":["Gernot Kristof, Elias Eulig, Marc Kachelrieß"]},"id":{"eki":["195019700X"],"doi":["10.1002/mp.18121"]},"relHost":[{"physDesc":[{"extent":"Online-Ressource"}],"name":{"displayForm":["American Association of Physicists in Medicine ; American Institute of Physics"]},"recId":"265784867","pubHistory":["1.1974 -"],"id":{"zdb":["1466421-5"],"eki":["265784867"],"issn":["2473-4209","1522-8541"]},"disp":"Noise-augmented deep denoising a method to boost CT image denoising networksMedical physics","part":{"volume":"52","extent":"13","pages":"1-13","text":"52(2025), 10 vom: Okt., Artikel-ID e18121, Seite 1-13","year":"2025","issue":"10"},"titleAlt":[{"title":"Medical physics online"}],"note":["Gesehen am 01.08.2025"],"origin":[{"dateIssuedDisp":"1974-","publisherPlace":"Hoboken, NJ ; College Park, Md. ; New York, NY","dateIssuedKey":"1974","publisher":"Wiley ; AAPM ; [Verlag nicht ermittelbar]"}],"title":[{"title_sort":"Medical physics","title":"Medical physics"}],"type":{"bibl":"periodical","media":"Online-Ressource"},"language":["eng"]}]} 
SRT |a KRISTOFGERNOISEAUGME2320