Measuring and Controlling the Spectral Bias for Self-Supervised Image Denoising

2025-09-30

AI Summary

Self-supervised denoising methods often suffer from insufficient preservation of high-frequency structural details in paired noisy images and are prone to overfitting high-frequency noise. Method: This paper proposes the Spectral Control Network (SCNet), which (1) introduces a band-selective strategy and an image spectral similarity metric to guide the network toward structurally relevant frequency bands; (2) designs a Spectral Separation and Reconstruction (SSR) module to explicitly decouple noise from high-frequency structural details in the frequency domain; and (3) imposes a Lipschitz constant constraint to suppress noise overfitting—enabling spectral response control without modifying the backbone architecture. Contribution/Results: Extensive experiments on both synthetic and real-world datasets demonstrate that SCNet significantly improves high-frequency structural fidelity, accelerates convergence, and reduces the tendency to learn noise. It establishes an interpretable and controllable frequency-domain modeling paradigm for self-supervised denoising.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Diffusion Models for Vision
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Responsible Web: Measurement, analysis, and circumvention of Web censorship
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Current self-supervised denoising methods for paired noisy images typically involve mapping one noisy image through the network to the other noisy image. However, after measuring the spectral bias of such methods using our proposed Image Pair Frequency-Band Similarity, it suffers from two practical limitations. Firstly, the high-frequency structural details in images are not preserved well enough. Secondly, during the process of fitting high frequencies, the network learns high-frequency noise from the mapped noisy images. To address these challenges, we introduce a Spectral Controlling network (SCNet) to optimize self-supervised denoising of paired noisy images. First, we propose a selection strategy to choose frequency band components for noisy images, to accelerate the convergence speed of training. Next, we present a parameter optimization method that restricts the learning ability of convolutional kernels to high-frequency noise using the Lipschitz constant, without changing the network structure. Finally, we introduce the Spectral Separation and low-rank Reconstruction module (SSR module), which separates noise and high-frequency details through frequency domain separation and low-rank space reconstruction, to retain the high-frequency structural details of images. Experiments performed on synthetic and real-world datasets verify the effectiveness of SCNet.

Problem

• Addresses poor preservation of high-frequency image details
• Reduces network learning of high-frequency noise artifacts
• Optimizes self-supervised denoising using spectral control methods

Innovation

• Frequency band selection strategy accelerates training convergence
• Lipschitz constant restricts kernels from learning high-frequency noise
• Spectral Separation and Reconstruction module preserves image details