A Master Class on Reproducibility: A Student Hackathon on Advanced MRI Reconstruction Methods
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912849506336768 |
|---|---|
| author | Felsner, Lina Kafali, Sevgi G. Eichhorn, Hannah Leth, Agnes A. J. Batvinskas, Aidas Datchev, Andre Klemm, Fabian Aulich, Jan Leepagorn, Puntika Klinger, Ruben Rueckert, Daniel Schnabel, Julia A. |
| author_facet | Felsner, Lina Kafali, Sevgi G. Eichhorn, Hannah Leth, Agnes A. J. Batvinskas, Aidas Datchev, Andre Klemm, Fabian Aulich, Jan Leepagorn, Puntika Klinger, Ruben Rueckert, Daniel Schnabel, Julia A. |
| contents | We report the design, protocol, and outcomes of a student reproducibility hackathon focused on replicating the results of three influential MRI reconstruction papers: (a) MoDL, an unrolled model-based network with learned denoising; (b) HUMUS-Net, a hybrid unrolled multiscale CNN+Transformer architecture; and (c) an untrained, physics-regularized dynamic MRI method that uses a quantitative MR model for early stopping. We describe the setup of the hackathon and present reproduction outcomes alongside additional experiments, and we detail fundamental practices for building reproducible codebases. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_18314 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | A Master Class on Reproducibility: A Student Hackathon on Advanced MRI Reconstruction Methods Felsner, Lina Kafali, Sevgi G. Eichhorn, Hannah Leth, Agnes A. J. Batvinskas, Aidas Datchev, Andre Klemm, Fabian Aulich, Jan Leepagorn, Puntika Klinger, Ruben Rueckert, Daniel Schnabel, Julia A. Machine Learning We report the design, protocol, and outcomes of a student reproducibility hackathon focused on replicating the results of three influential MRI reconstruction papers: (a) MoDL, an unrolled model-based network with learned denoising; (b) HUMUS-Net, a hybrid unrolled multiscale CNN+Transformer architecture; and (c) an untrained, physics-regularized dynamic MRI method that uses a quantitative MR model for early stopping. We describe the setup of the hackathon and present reproduction outcomes alongside additional experiments, and we detail fundamental practices for building reproducible codebases. |
| title | A Master Class on Reproducibility: A Student Hackathon on Advanced MRI Reconstruction Methods |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2601.18314 |