IDF: Iterative Dynamic Filtering Networks for Generalizable Image Denoising

TL;DR

This paper introduces IDF, a novel image denoising method using iterative dynamic filtering with locally adaptive kernels, achieving high generalization to unseen noise types and levels with a compact model.

Contribution

The paper proposes a new iterative dynamic filtering network that generates pixel-wise kernels for robust, generalizable image denoising, even trained on a single noise level.

Findings

Outperforms existing methods on diverse noise types and levels.

Uses a compact model (~0.04M) with efficient operations.

Demonstrates strong generalization despite training on single Gaussian noise.

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.