1. [Publications](/index.php/publications)
2. Noise2Noise: Learning Image Restoration without Clean Data
 
 # Noise2Noise: Learning Image Restoration without Clean Data

  ![](/sites/default/files/styles/wide/public/publications/n2n-representative_0.png?itok=Zt11Dkr5)

 We apply basic statistical reasoning to signal reconstruction by machine learning — learning to map corrupted observations to clean signals — with a simple and powerful conclusion: under certain common circumstances, it is possible to learn to restore signals without ever observing clean ones, at performance close or equal to training using clean exemplars. We show applications in photographic noise removal, denoising of synthetic Monte Carlo images, and reconstruction of MRI scans from undersampled inputs, all based on only observing corrupted data.



 ## Authors



Jaakko Lehtinen (NVIDIA & Aalto University)

[Jacob Munkberg](/index.php/person/jacob-munkberg)

[Jon Hasselgren](/index.php/person/jon-hasselgren)

[Samuli Laine](/index.php/person/samuli-laine)

[Tero Karras](/index.php/person/tero-karras)

Miika Aittala (MIT CSAIL)

[Timo Aila](/index.php/person/timo-aila)

 

 

 ## Publication Date



Sunday, July 15, 2018

 

 ## Published in



[Proc. ICML 2018](https://icml.cc/)

 

 ## Research Area



[Computational Photography and Imaging](/index.php/research-area/computational-photography-imaging)

[Computer Vision](/index.php/research-area/computer-vision)

[Artificial Intelligence and Machine Learning ](/index.php/research-area/machine-learning-artificial-intelligence)

 

 

 ## External Links



[Published paper (PMLR, Open Access)](http://proceedings.mlr.press/v80/lehtinen18a.html)

[Final published version (arXiv)](https://arxiv.org/abs/1803.04189)