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

TL;DR

This paper introduces Next-Scale Prediction, a self-supervised method for real-world image denoising that effectively balances noise removal and detail preservation by using cross-scale training pairs.

Contribution

It proposes a novel self-supervised paradigm that decouples noise decorrelation from detail preservation, enabling state-of-the-art denoising and super-resolution without retraining.

Findings

Achieves state-of-the-art performance on real-world benchmarks.

Supports super-resolution of noisy images without retraining.

Effectively balances noise removal and detail preservation.

Abstract

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.