User: Guest  Login
Title:

Test-Time Training Can Close the Natural Distribution Shift Performance Gap in Deep Learning Based Compressed Sensing

Document type:
Konferenzbeitrag
Author(s):
Darestani, Mohammad Zalbagi; Liu, Jiayu; Heckel, Reinhard
Abstract:
Deep learning based image reconstruction methods outperform traditional methods. However, neural networks suffer from a performance drop when applied to images from a different distribution than the training images. For example, a model trained for reconstructing knees in accelerated magnetic resonance imaging (MRI) does not reconstruct brains well, even though the same network trained on brains reconstructs brains perfectly well. Thus there is a distribution shift performance gap for a given ne...     »
Editor:
Chaudhuri, Kamalika; Jegelka, Stefanie; Song, Le; Szepesvari, Csaba; Niu, Gang; Sabato, Sivan
Book / Congress title:
Proceedings of the 39th International Conference on Machine Learning
Volume:
162
Publisher:
PMLR
Year:
2022
Month:
17--23 Jul
Pages:
4754--4776
Bookseries title:
Proceedings of Machine Learning Research
WWW:
https://proceedings.mlr.press/v162/darestani22a.html
 BibTeX