Metraipolis › archive › dc_783ca5047246
Noise2Noise Revisited: Training Pair Distributions Dominate Loss Choice in Self-Supervised Denoising — arXiv:2609.16788
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.16788
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T22:40:12.237Z
- SHA-256 of the archived text
5fe3169100f411793f1d1f3862223a4ae836fa880667bd832c3e6de4b453ae6c- Stable URL of this page
https://metraipolis.com/documento/dc_783ca5047246
How to cite
Noise2Noise Revisited: Training Pair Distributions Dominate Loss Choice in Self-Supervised Denoising — arXiv:2609.16788 — arXiv (en), https://arxiv.org/abs/2609.16788 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_783ca5047246 (Metraipolis archive dc_783ca5047246, sha256 5fe3169100f411793f1d1f3862223a4ae836fa880667bd832c3e6de4b453ae6c)
BibTeX
@misc{dc783ca5047246,
title = {Noise2Noise Revisited: Training Pair Distributions Dominate Loss Choice in Self-Supervised Denoising — arXiv:2609.16788},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.16788},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_783ca5047246; sha256 5fe3169100f411793f1d1f3862223a4ae836fa880667bd832c3e6de4b453ae6c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.16788. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Noise2Noise (N2N) trains denoisers on pairs of independently corrupted observations, eliminating clean references. We stress-test two natural conjectures about why the L1 loss outperforms L2 here. First, the hypothesis that the L1 loss confers robustness via parameter sparsity confuses the loss with Lasso regularization: an explicit Lasso penalty produces the predicted sparsity yet fails to reproduce L1's cross-noise behavior, while L1- and L2-trained weight distributions are indistinguishable. Second, the population optima of the two losses coincide exactly for symmetric signal posteriors and nearly so for concentrated ones. Measured differences are therefore dominated by optimization dynamics (bounded-influence gradients), which we probe with gradient statistics and contaminated-target training. On Kodak24 with five synthetic noise families, the L1 loss holds a statistically significant edge over L2, below 1 dB PSNR, holding across three seeds on 13 of the 14 noise columns. On real camera noise the loss is not the decisive variable in distribution: on official SIDD validation blocks, synthetic-Gaussian-trained N2N models gain only 0.8 to 3.7 dB over the noisy input regardless of loss, while retraining on SIDD's own noisy pairs, never reading ground truth, gains 9.4 to 11.0 dB, far ahead of BM3D. All metrics are on raw network outputs, and the study makes no leaderboard claim. The training pair distribution, not the loss, carries the inductive bias. That design rule applies wherever clean references are unobtainable, from microscopy to industrial inspection sensors. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-16 in eess.IV). Categorie arXiv: Machine Learning (cs.LG) ; Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV). Pagina del lavoro: https://arxiv.org/abs/2609.16788 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-15T22:18:16.509Z dalla pagina https://arxiv.org/catchup/eess.IV/2026-09-16?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).