Home > Publications database > Reconstructing and analyzing the invariances of low-dose CT image denoising networks. > print

001     293809
005     20250107162137.0
024 7 _ |a 10.1002/mp.17413
|2 doi
024 7 _ |a pmid:39348044
|2 pmid
024 7 _ |a 0094-2405
|2 ISSN
024 7 _ |a 1522-8541
|2 ISSN
024 7 _ |a 2473-4209
|2 ISSN
037 _ _ |a DKFZ-2024-01967
041 _ _ |a English
082 _ _ |a 610
100 1 _ |a Eulig, Elias
|0 P:(DE-He78)9f72962d56560b57e5630885c8f8b31d
|b 0
|e First author
|u dkfz
245 _ _ |a Reconstructing and analyzing the invariances of low-dose CT image denoising networks.
260 _ _ |a College Park, Md.
|c 2025
|b AAPM
336 7 _ |a article
|2 DRIVER
336 7 _ |a Output Types/Journal article
|2 DataCite
336 7 _ |a Journal Article
|b journal
|m journal
|0 PUB:(DE-HGF)16
|s 1736263248_4144
|2 PUB:(DE-HGF)
336 7 _ |a ARTICLE
|2 BibTeX
336 7 _ |a JOURNAL_ARTICLE
|2 ORCID
336 7 _ |a Journal Article
|0 0
|2 EndNote
500 _ _ |a #EA:E025#LA:E025# / 2025 Jan;52(1):188-200
520 _ _ |a 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.To improve the interpretability of deep learning-based low-dose CT image denoising networks.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.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.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.
536 _ _ |a 315 - Bildgebung und Radioonkologie (POF4-315)
|0 G:(DE-HGF)POF4-315
|c POF4-315
|f POF IV
|x 0
588 _ _ |a Dataset connected to CrossRef, PubMed, , Journals: inrepo02.dkfz.de
650 _ 7 |a computed tomography
|2 Other
650 _ 7 |a deep learning
|2 Other
650 _ 7 |a explainability
|2 Other
650 _ 7 |a invariances
|2 Other
650 _ 7 |a low‐dose
|2 Other
650 _ 7 |a robustness
|2 Other
700 1 _ |a Jäger, Fabian
|0 P:(DE-He78)ceed8085bb2faf0ec822fe89e8a085d3
|b 1
|u dkfz
700 1 _ |a Maier, Joscha
|0 P:(DE-He78)3c462b1378ce0906e7320c94e514abfa
|b 2
|u dkfz
700 1 _ |a Ommer, Björn
|b 3
700 1 _ |a Kachelriess, Marc
|0 P:(DE-He78)f288a8f92f092ddb41d52b1aeb915323
|b 4
|e Last author
|u dkfz
773 _ _ |a 10.1002/mp.17413
|g p. mp.17413
|0 PERI:(DE-600)1466421-5
|n 1
|p 188-200
|t Medical physics
|v 52
|y 2025
|x 0094-2405
909 C O |p VDB
|o oai:inrepo02.dkfz.de:293809
910 1 _ |a Deutsches Krebsforschungszentrum
|0 I:(DE-588b)2036810-0
|k DKFZ
|b 0
|6 P:(DE-He78)9f72962d56560b57e5630885c8f8b31d
910 1 _ |a Deutsches Krebsforschungszentrum
|0 I:(DE-588b)2036810-0
|k DKFZ
|b 1
|6 P:(DE-He78)ceed8085bb2faf0ec822fe89e8a085d3
910 1 _ |a Deutsches Krebsforschungszentrum
|0 I:(DE-588b)2036810-0
|k DKFZ
|b 2
|6 P:(DE-He78)3c462b1378ce0906e7320c94e514abfa
910 1 _ |a Deutsches Krebsforschungszentrum
|0 I:(DE-588b)2036810-0
|k DKFZ
|b 4
|6 P:(DE-He78)f288a8f92f092ddb41d52b1aeb915323
913 1 _ |a DE-HGF
|b Gesundheit
|l Krebsforschung
|1 G:(DE-HGF)POF4-310
|0 G:(DE-HGF)POF4-315
|3 G:(DE-HGF)POF4
|2 G:(DE-HGF)POF4-300
|4 G:(DE-HGF)POF
|v Bildgebung und Radioonkologie
|x 0
914 1 _ |y 2024
915 _ _ |a DEAL Wiley
|0 StatID:(DE-HGF)3001
|2 StatID
|d 2023-10-21
|w ger
915 _ _ |a JCR
|0 StatID:(DE-HGF)0100
|2 StatID
|b MED PHYS : 2022
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)0200
|2 StatID
|b SCOPUS
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)0300
|2 StatID
|b Medline
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)0320
|2 StatID
|b PubMed Central
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)0600
|2 StatID
|b Ebsco Academic Search
|d 2023-10-21
915 _ _ |a Peer Review
|0 StatID:(DE-HGF)0030
|2 StatID
|b ASC
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)0199
|2 StatID
|b Clarivate Analytics Master Journal List
|d 2023-10-21
915 _ _ |a WoS
|0 StatID:(DE-HGF)0113
|2 StatID
|b Science Citation Index Expanded
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)0150
|2 StatID
|b Web of Science Core Collection
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)1030
|2 StatID
|b Current Contents - Life Sciences
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)0160
|2 StatID
|b Essential Science Indicators
|d 2023-10-21
915 _ _ |a DBCoverage
|0 StatID:(DE-HGF)1110
|2 StatID
|b Current Contents - Clinical Medicine
|d 2023-10-21
915 _ _ |a IF < 5
|0 StatID:(DE-HGF)9900
|2 StatID
|d 2023-10-21
920 2 _ |0 I:(DE-He78)E025-20160331
|k E025
|l E025 Röntgenbildgebung und Computertomographie
|x 0
920 1 _ |0 I:(DE-He78)E025-20160331
|k E025
|l E025 Röntgenbildgebung und Computertomographie
|x 0
920 0 _ |0 I:(DE-He78)E025-20160331
|k E025
|l E025 Röntgenbildgebung und Computertomographie
|x 0
980 _ _ |a journal
980 _ _ |a VDB
980 _ _ |a I:(DE-He78)E025-20160331
980 _ _ |a UNRESTRICTED

LibraryCollectionCLSMajorCLSMinorLanguageAuthor
Marc 21