Reconstructing and analyzing the invariances of low-dose CT image denoising networks
Background Deep learning-based methods led to significant advancements in many areas of medical imaging, most of which are concerned with the reduction of artifacts caused by motion, scatter, or noise. However, with most neural networks being black boxes, they remain notoriously difficult to interpr...
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Dokumenttyp: | Article (Journal) |
| Sprache: | Englisch |
| Veröffentlicht: |
January 2025
|
| In: |
Medical physics
Year: 2025, Jahrgang: 52, Heft: 1, Pages: 188-200 |
| ISSN: | 2473-4209 |
| DOI: | 10.1002/mp.17413 |
| Online-Zugang: | Verlag, kostenfrei, Volltext: https://doi.org/10.1002/mp.17413 Verlag, kostenfrei, Volltext: https://onlinelibrary.wiley.com/doi/abs/10.1002/mp.17413 
|
| Verfasserangaben: | Elias Eulig, Fabian Jäger, Joscha Maier, Björn Ommer, Marc Kachelrieß |
MARC
| LEADER | 00000caa a2200000 c 4500 | ||
|---|---|---|---|
| 001 | 1919847294 | ||
| 003 | DE-627 | ||
| 005 | 20250717001326.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 250314s2025 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1002/mp.17413 |2 doi | |
| 035 | |a (DE-627)1919847294 | ||
| 035 | |a (DE-599)KXP1919847294 | ||
| 035 | |a (OCoLC)1528043180 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 33 |2 sdnb | ||
| 100 | 1 | |a Eulig, Elias |d 1995- |e VerfasserIn |0 (DE-588)1191843181 |0 (DE-627)1670315185 |4 aut | |
| 245 | 1 | 0 | |a Reconstructing and analyzing the invariances of low-dose CT image denoising networks |c Elias Eulig, Fabian Jäger, Joscha Maier, Björn Ommer, Marc Kachelrieß |
| 264 | 1 | |c January 2025 | |
| 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 14.03.2025 | ||
| 520 | |a Background Deep learning-based methods led to significant advancements in many areas of medical imaging, most of which are concerned with the reduction of artifacts caused by motion, scatter, or noise. However, with most neural networks being black boxes, they remain notoriously difficult to interpret, hindering their clinical implementation. In particular, it has been shown that networks exhibit invariances w.r.t. input features, that is, they learn to ignore certain information in the input data. Purpose To improve the interpretability of deep learning-based low-dose CT image denoising networks. Methods We learn a complete data representation of low-dose input images using a conditional variational autoencoder (cVAE). In this representation, invariances of any given denoising network are then disentangled from the information it is not invariant to using a conditional invertible neural network (cINN). At test time, image-space invariances are generated by applying the inverse of the cINN and subsequent decoding using the cVAE. We propose two methods to analyze sampled invariances and to find those that correspond to alterations of anatomical structures. Results The proposed method is applied to four popular deep learning-based low-dose CT image denoising networks. We find that the networks are not only invariant to noise amplitude and realizations, but also to anatomical structures. Conclusions The proposed method is capable of reconstructing and analyzing invariances of deep learning-based low-dose CT image denoising networks. This is an important step toward interpreting deep learning-based methods for medical imaging, which is essential for their clinical implementation. | ||
| 650 | 4 | |a computed tomography | |
| 650 | 4 | |a deep learning | |
| 650 | 4 | |a explainability | |
| 650 | 4 | |a invariances | |
| 650 | 4 | |a low-dose | |
| 650 | 4 | |a robustness | |
| 700 | 1 | |a Jäger, Fabian |d 1996- |e VerfasserIn |0 (DE-588)1358990123 |0 (DE-627)1919317783 |4 aut | |
| 700 | 1 | |a Maier, Joscha |d 1988- |e VerfasserIn |0 (DE-588)1185600868 |0 (DE-627)1664987231 |4 aut | |
| 700 | 1 | |a Ommer, Björn |d 1981- |e VerfasserIn |0 (DE-588)1034893106 |0 (DE-627)746457510 |0 (DE-576)382507916 |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), 1 vom: Jan., Seite 188-200 |h Online-Ressource |w (DE-627)265784867 |w (DE-600)1466421-5 |w (DE-576)074891243 |x 2473-4209 |7 nnas |a Reconstructing and analyzing the invariances of low-dose CT image denoising networks |
| 773 | 1 | 8 | |g volume:52 |g year:2025 |g number:1 |g month:01 |g pages:188-200 |g extent:13 |a Reconstructing and analyzing the invariances of low-dose CT image denoising networks |
| 856 | 4 | 0 | |u https://doi.org/10.1002/mp.17413 |x Verlag |x Resolving-System |z kostenfrei |3 Volltext |
| 856 | 4 | 0 | |u https://onlinelibrary.wiley.com/doi/abs/10.1002/mp.17413 |x Verlag |z kostenfrei |3 Volltext |
| 951 | |a AR | ||
| 992 | |a 20250314 | ||
| 993 | |a Article | ||
| 994 | |a 2025 | ||
| 998 | |g 120866544 |a Kachelrieß, Marc |m 120866544:Kachelrieß, Marc |d 50000 |e 50000PK120866544 |k 0/50000/ |p 5 | ||
| 998 | |g 1358990123 |a Jäger, Fabian |m 1358990123:Jäger, Fabian |d 130000 |e 130000PJ1358990123 |k 0/130000/ |p 2 | ||
| 998 | |g 1191843181 |a Eulig, Elias |m 1191843181:Eulig, Elias |d 130000 |e 130000PE1191843181 |k 0/130000/ |p 1 |x j | ||
| 999 | |a KXP-PPN1919847294 |e 4687171605 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"origin":[{"dateIssuedKey":"2025","dateIssuedDisp":"January 2025"}],"language":["eng"],"relHost":[{"note":["Gesehen am 01.08.2025"],"physDesc":[{"extent":"Online-Ressource"}],"disp":"Reconstructing and analyzing the invariances of low-dose CT image denoising networksMedical physics","language":["eng"],"origin":[{"publisher":"Wiley ; AAPM ; [Verlag nicht ermittelbar]","dateIssuedDisp":"1974-","publisherPlace":"Hoboken, NJ ; College Park, Md. ; New York, NY","dateIssuedKey":"1974"}],"type":{"bibl":"periodical","media":"Online-Ressource"},"part":{"text":"52(2025), 1 vom: Jan., Seite 188-200","extent":"13","volume":"52","issue":"1","pages":"188-200","year":"2025"},"titleAlt":[{"title":"Medical physics online"}],"title":[{"title":"Medical physics","title_sort":"Medical physics"}],"id":{"zdb":["1466421-5"],"eki":["265784867"],"issn":["2473-4209","1522-8541"]},"recId":"265784867","pubHistory":["1.1974 -"],"name":{"displayForm":["American Association of Physicists in Medicine ; American Institute of Physics"]}}],"physDesc":[{"extent":"13 S."}],"note":["Gesehen am 14.03.2025"],"name":{"displayForm":["Elias Eulig, Fabian Jäger, Joscha Maier, Björn Ommer, Marc Kachelrieß"]},"id":{"eki":["1919847294"],"doi":["10.1002/mp.17413"]},"recId":"1919847294","title":[{"title":"Reconstructing and analyzing the invariances of low-dose CT image denoising networks","title_sort":"Reconstructing and analyzing the invariances of low-dose CT image denoising networks"}],"type":{"bibl":"article-journal","media":"Online-Ressource"},"person":[{"role":"aut","family":"Eulig","given":"Elias","roleDisplay":"VerfasserIn","display":"Eulig, Elias"},{"family":"Jäger","roleDisplay":"VerfasserIn","given":"Fabian","display":"Jäger, Fabian","role":"aut"},{"family":"Maier","display":"Maier, Joscha","roleDisplay":"VerfasserIn","given":"Joscha","role":"aut"},{"roleDisplay":"VerfasserIn","given":"Björn","display":"Ommer, Björn","family":"Ommer","role":"aut"},{"role":"aut","display":"Kachelrieß, Marc","roleDisplay":"VerfasserIn","given":"Marc","family":"Kachelrieß"}]} | ||
| SRT | |a EULIGELIASRECONSTRUC2025 | ||