mlsanityData to 7 Oct 2026

Next-Scale Prediction: A Self-Supervised Approach for Real-World Image Denoising

arXiv 2512.21038 ↗First posted 24 Dec 2025cs.CVcited by 1

Not on the map. Why?

Abstract (arXiv)

Self-supervised real-world image denoising remains a fundamental challenge, arising from the antagonistic trade-off between decorrelating spatially structured noise and preserving high-frequency details. Existing blind-spot network (BSN) methods rely on pixel-shuffle downsampling (PD) to decorrelate noise, but aggressive downsampling fragments fine structures, while milder downsampling fails to remove correlated noise. To address this, we introduce Next-Scale Prediction (NSP), a novel self-supervised paradigm that decouples noise decorrelation from detail preservation. NSP constructs cross-scale training pairs, where BSN takes low-resolution, fully decorrelated sub-images as input to predict high-resolution targets that retain fine details. As a by-product, NSP naturally supports super-resolution of noisy images without retraining or modification. Extensive experiments demonstrate that NSP achieves state-of-the-art self-supervised denoising performance on real-world benchmarks, significantly alleviating the long-standing conflict between noise decorrelation and detail preservation. The code is available at https://github.com/XLearning-SCU/2026-CVPR-NSP.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 33 of its 43 references that are in the map. First posted on arXiv 24 Dec 2025.

Builds on

33 of 43 references.

References · 33

  1. 2017Attention Is All You Needcited by 59,932
  2. 2016Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoisingcited by 1,352
  3. 2024Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Predictioncited by 1,005
  4. 2018Noise2Noise: Learning Image Restoration without Clean Datacited by 526
  5. 2017FFDNet: Toward a Fast and Flexible Solution for CNN based Image Denoisingcited by 509
  6. 2018Noise2Void - Learning Denoising from Single Noisy Imagescited by 359
  7. 2017"Zero-Shot" Super-Resolution using Deep Internal Learningcited by 285
  8. 2018Toward Convolutional Blind Denoising of Real Photographscited by 274
  9. 2019Noise2Self: Blind Denoising by Self-Supervisioncited by 274
  10. 2018Unprocessing Images for Learned Raw Denoisingcited by 250
  11. 2015Trainable Nonlinear Reaction Diffusion: A Flexible Framework for Fast and Effective Image Restorationcited by 194
  12. 2019Real Image Denoising with Feature Attentioncited by 176
21 more
  1. 2021Neighbor2Neighbor: Self-Supervised Denoising from Single Noisy Imagescited by 171
  2. 2019High-Quality Self-Supervised Deep Image Denoisingcited by 161
  3. 2019Deep Learning on Image Denoising: An overviewcited by 128
  4. 2019Noisier2Noise: Learning to Denoise from Unpaired Noisy Datacited by 125
  5. 2019Variational Denoising Network: Toward Blind Noise Modeling and Removalcited by 117
  6. 2020Meta-Transfer Learning for Zero-Shot Super-Resolutioncited by 116
  7. 2022AP-BSN: Self-Supervised Denoising for Real-World Images via Asymmetric PD and Blind-Spot Networkcited by 92
  8. 2022Blind2Unblind: Self-Supervised Image Denoising with Visible Blind Spotscited by 83
  9. 2020Unpaired Learning of Deep Image Denoisingcited by 74
  10. 2022C2N: Practical Generative Noise Modeling for Real-World Denoisingcited by 56
  11. 2019When AWGN-based Denoiser Meets Real Noisescited by 53
  12. 2022CVF-SID: Cyclic multi-Variate Function for Self-Supervised Image Denoising by Disentangling Noise from Imagecited by 45
  13. 2023LG-BPN: Local and Global Blind-Patch Network for Self-Supervised Real-World Denoisingcited by 37
  14. 2022Spatial-Spectral Transformer for Hyperspectral Image Denoisingcited by 34
  15. 2023Spatially Adaptive Self-Supervised Learning for Real-World Image Denoisingcited by 33
  16. 2023Random Sub-Samples Generation for Self-Supervised Real Image Denoisingcited by 31
  17. 2022Multi-Scale Adaptive Network for Single Image Denoisingcited by 17
  18. 2023Exploring Efficient Asymmetric Blind-Spots for Self-Supervised Denoising in Real-World Scenarioscited by 17
  19. 2023Self-supervised Image Denoising with Downsampled Invariance Loss and Conditional Blind-Spot Networkcited by 12
  20. 2024Rethinking Transformer-Based Blind-Spot Network for Self-Supervised Image Denoisingcited by 9
  21. 2022RZSR: Reference-based Zero-Shot Super-Resolution with Depth Guided Self-Exemplarscited by 8

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Noise2Noise Revisited: Training Pair Distributions Dominate Loss Choice in Self-Supervised Denoisingcited by 0