Posterior temperature optimized Bayesian models for inverse problems in medical imaging
We present Posterior Temperature Optimized Bayesian Inverse Models (POTOBIM), an unsupervised Bayesian approach to inverse problems in medical imaging using mean-field variational inference with a fully tempered posterior. Bayesian methods exhibit useful properties for approaching inverse tasks, suc...
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Dokumenttyp: | Article (Journal) |
| Sprache: | Englisch |
| Veröffentlicht: |
May 2022
|
| In: |
Medical image analysis
Year: 2022, Jahrgang: 78, Pages: 1-12 |
| ISSN: | 1361-8423 |
| DOI: | 10.1016/j.media.2022.102382 |
| Online-Zugang: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1016/j.media.2022.102382 Verlag, lizenzpflichtig, Volltext: https://www.sciencedirect.com/science/article/pii/S1361841522000342 
|
| Verfasserangaben: | Max-Heinrich Laves, Malte Tölle, Alexander Schlaefer, Sandy Engelhardt |
MARC
| LEADER | 00000caa a2200000 c 4500 | ||
|---|---|---|---|
| 001 | 1806875764 | ||
| 003 | DE-627 | ||
| 005 | 20230427160805.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 220613s2022 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1016/j.media.2022.102382 |2 doi | |
| 035 | |a (DE-627)1806875764 | ||
| 035 | |a (DE-599)KXP1806875764 | ||
| 035 | |a (OCoLC)1341461332 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 28 |2 sdnb | ||
| 100 | 1 | |a Laves, Max-Heinrich |d 1988- |e VerfasserIn |0 (DE-588)1239905637 |0 (DE-627)1767966962 |4 aut | |
| 245 | 1 | 0 | |a Posterior temperature optimized Bayesian models for inverse problems in medical imaging |c Max-Heinrich Laves, Malte Tölle, Alexander Schlaefer, Sandy Engelhardt |
| 264 | 1 | |c May 2022 | |
| 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 Available online 11 February 2022 | ||
| 500 | |a Gesehen am 13.06.2022 | ||
| 520 | |a We present Posterior Temperature Optimized Bayesian Inverse Models (POTOBIM), an unsupervised Bayesian approach to inverse problems in medical imaging using mean-field variational inference with a fully tempered posterior. Bayesian methods exhibit useful properties for approaching inverse tasks, such as tomographic reconstruction or image denoising. A suitable prior distribution introduces regularization, which is needed to solve the ill-posed problem and reduces overfitting the data. In practice, however, this often results in a suboptimal posterior temperature, and the full potential of the Bayesian approach is not being exploited. In POTOBIM, we optimize both the parameters of the prior distribution and the posterior temperature with respect to reconstruction accuracy using Bayesian optimization with Gaussian process regression. Our method is extensively evaluated on four different inverse tasks on a variety of modalities with images from public data sets and we demonstrate that an optimized posterior temperature outperforms both non-Bayesian and Bayesian approaches without temperature optimization. The use of an optimized prior distribution and posterior temperature leads to improved accuracy and uncertainty estimation and we show that it is sufficient to find these hyperparameters per task domain. Well-tempered posteriors yield calibrated uncertainty, which increases the reliability in the predictions. Our source code is publicly available at github.com/Cardio-AI/mfvi-dip-mia. | ||
| 650 | 4 | |a Deep learning | |
| 650 | 4 | |a Hallucination | |
| 650 | 4 | |a Variational inference | |
| 700 | 1 | |a Tölle, Malte |e VerfasserIn |0 (DE-588)1259770931 |0 (DE-627)180687444X |4 aut | |
| 700 | 1 | |a Schlaefer, Alexander |d 1975- |e VerfasserIn |0 (DE-588)123043522 |0 (DE-627)566664453 |0 (DE-576)185081134 |4 aut | |
| 700 | 1 | |a Engelhardt, Sandy |d 1987- |e VerfasserIn |0 (DE-588)1122674465 |0 (DE-627)876003080 |0 (DE-576)481436049 |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t Medical image analysis |d Amsterdam [u.a.] : Elsevier Science, 1996 |g 78(2022), Artikel-ID 102382, 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 Posterior temperature optimized Bayesian models for inverse problems in medical imaging |
| 773 | 1 | 8 | |g volume:78 |g year:2022 |g elocationid:102382 |g pages:1-12 |g extent:12 |a Posterior temperature optimized Bayesian models for inverse problems in medical imaging |
| 856 | 4 | 0 | |u https://doi.org/10.1016/j.media.2022.102382 |x Verlag |x Resolving-System |z lizenzpflichtig |3 Volltext |
| 856 | 4 | 0 | |u https://www.sciencedirect.com/science/article/pii/S1361841522000342 |x Verlag |z lizenzpflichtig |3 Volltext |
| 951 | |a AR | ||
| 992 | |a 20220613 | ||
| 993 | |a Article | ||
| 994 | |a 2022 | ||
| 998 | |g 1122674465 |a Engelhardt, Sandy |m 1122674465:Engelhardt, Sandy |d 910000 |d 910100 |e 910000PE1122674465 |e 910100PE1122674465 |k 0/910000/ |k 1/910000/910100/ |p 4 |y j | ||
| 998 | |g 1259770931 |a Tölle, Malte |m 1259770931:Tölle, Malte |d 910000 |d 910100 |e 910000PT1259770931 |e 910100PT1259770931 |k 0/910000/ |k 1/910000/910100/ |p 2 | ||
| 999 | |a KXP-PPN1806875764 |e 4149066922 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"relHost":[{"part":{"text":"78(2022), Artikel-ID 102382, Seite 1-12","extent":"12","pages":"1-12","volume":"78","year":"2022"},"language":["eng"],"note":["Gesehen am 16.05.23"],"pubHistory":["1.1996/97 -"],"type":{"bibl":"periodical","media":"Online-Ressource"},"disp":"Posterior temperature optimized Bayesian models for inverse problems in medical imagingMedical image analysis","titleAlt":[{"title":"Medical image analysis online"}],"recId":"306365081","physDesc":[{"extent":"Online-Ressource"}],"id":{"zdb":["1497450-2"],"issn":["1361-8423"],"eki":["306365081"]},"title":[{"title_sort":"Medical image analysis","title":"Medical image analysis"}],"origin":[{"publisherPlace":"Amsterdam [u.a.]","dateIssuedKey":"1996","dateIssuedDisp":"1996-","publisher":"Elsevier Science"}]}],"origin":[{"dateIssuedKey":"2022","dateIssuedDisp":"May 2022"}],"type":{"bibl":"article-journal","media":"Online-Ressource"},"title":[{"title_sort":"Posterior temperature optimized Bayesian models for inverse problems in medical imaging","title":"Posterior temperature optimized Bayesian models for inverse problems in medical imaging"}],"id":{"eki":["1806875764"],"doi":["10.1016/j.media.2022.102382"]},"physDesc":[{"extent":"12 S."}],"person":[{"display":"Laves, Max-Heinrich","given":"Max-Heinrich","family":"Laves","role":"aut"},{"family":"Tölle","role":"aut","display":"Tölle, Malte","given":"Malte"},{"given":"Alexander","display":"Schlaefer, Alexander","role":"aut","family":"Schlaefer"},{"given":"Sandy","display":"Engelhardt, Sandy","family":"Engelhardt","role":"aut"}],"recId":"1806875764","note":["Available online 11 February 2022","Gesehen am 13.06.2022"],"language":["eng"],"name":{"displayForm":["Max-Heinrich Laves, Malte Tölle, Alexander Schlaefer, Sandy Engelhardt"]}} | ||
| SRT | |a LAVESMAXHEPOSTERIORT2022 | ||