Zero-Shot Noise2Noise: Efficient Image Denoising Without Any Data

MCML Authors

Youssef Mansour

* Former Member

→ Group Reinhard Heckel
 Machine Learning and Information Processing

Reinhard Heckel

Prof. Dr.

Core PI

Machine Learning and Information Processing

Abstract

Recently, self-supervised neural networks have shown excellent image denoising performance. How-ever, current dataset free methods are either computationally expensive, require a noise model, or have inad-equate image quality. In this work we show that a simple 2-layer network, without any training data or knowledge of the noise distribution, can enable high-quality image denoising at low computational cost. Our approach is motivated by Noise2Noise and Neighbor2Neighbor and works well for denoising pixel-wise independent noise. Our experiments on artificial, real-world cam-era, and microscope noise show that our method termed ZS-N2N (Zero Shot Noise2Noise) often outperforms ex-isting dataset-free methods at a reduced cost, making it suitable for use cases with scarce data availability and limited compute.

inproceedings MH23

---

CVPR 2023

IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver, Canada, Jun 18-23, 2023.

---

Authors

Y. Mansour • R. Heckel

---

Links

DOI

---

Research Area

 A2 | Mathematical Foundations