IDF: Iterative Dynamic Filtering Networks for Generalizable Image Denoising

2025-08-27

AI Summary

Existing deep learning-based image denoising methods suffer from poor generalization to unseen noise types and intensities, are prone to overfitting, and heavily rely on large-scale labeled datasets and high computational resources. To address these limitations, this paper proposes a lightweight iterative dynamic filtering framework trained exclusively on single-level Gaussian noise, yet capable of generalizing effectively to diverse complex noise distributions—including Poisson, impulse, and mixed noise. The core innovation lies in a pixel-wise adaptive convolutional kernel generation mechanism that jointly integrates feature extraction, global statistical modeling, and local correlation modeling; a dedicated Kernel Prediction module dynamically synthesizes spatially varying filters tailored to local image structures. With only 0.04 million parameters, the model achieves state-of-the-art performance across multiple noise scenarios, demonstrating superior generalization, robustness, and computational efficiency.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Generation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Image denoising is a fundamental challenge in computer vision, with applications in photography and medical imaging. While deep learning-based methods have shown remarkable success, their reliance on specific noise distributions limits generalization to unseen noise types and levels. Existing approaches attempt to address this with extensive training data and high computational resources but they still suffer from overfitting. To address these issues, we conduct image denoising by utilizing dynamically generated kernels via efficient operations. This approach helps prevent overfitting and improves resilience to unseen noise. Specifically, our method leverages a Feature Extraction Module for robust noise-invariant features, Global Statistics and Local Correlation Modules to capture comprehensive noise characteristics and structural correlations. The Kernel Prediction Module then employs these cues to produce pixel-wise varying kernels adapted to local structures, which are then applied iteratively for denoising. This ensures both efficiency and superior restoration quality. Despite being trained on single-level Gaussian noise, our compact model (~ 0.04 M) excels across diverse noise types and levels, demonstrating the promise of iterative dynamic filtering for practical image denoising.

Problem

• Generalizing image denoising to unseen noise types
• Overcoming overfitting in deep learning denoising methods
• Reducing computational resources for robust noise removal

Innovation

• Iterative dynamic filtering with pixel-wise kernels
• Feature extraction capturing noise-invariant characteristics
• Compact model generalizing across diverse noise types