CDLNet: Noise-Adaptive Convolutional Dictionary Learning Network for Blind Denoising and Demosaicing

TL;DR

CDLNet is an interpretable, unrolled convolutional dictionary learning network that achieves state-of-the-art blind denoising and joint denoising and demosaicing performance, with strong generalization to unseen noise levels.

Contribution

This work introduces CDLNet, a novel unrolled convolutional dictionary learning network with noise-adaptive parameters, enhancing interpretability and performance in blind denoising and JDD tasks.

Findings

Outperforms state-of-the-art models at similar parameter counts.

Achieves near-perfect generalization to unseen noise levels.

Extends performance to joint denoising and demosaicing, including unsupervised learning.

Abstract

Deep learning based methods hold state-of-the-art results in low-level image processing tasks, but remain difficult to interpret due to their black-box construction. Unrolled optimization networks present an interpretable alternative to constructing deep neural networks by deriving their architecture from classical iterative optimization methods without use of tricks from the standard deep learning tool-box. So far, such methods have demonstrated performance close to that of state-of-the-art models while using their interpretable construction to achieve a comparably low learned parameter count. In this work, we propose an unrolled convolutional dictionary learning network (CDLNet) and demonstrate its competitive denoising and joint denoising and demosaicing (JDD) performance both in low and high parameter count regimes. Specifically, we show that the proposed model outperforms state-of-the-art fully convolutional denoising and JDD models when scaled to a similar parameter count. In addition, we leverage the model's interpretable construction to propose a noise-adaptive parameterization of thresholds in the network that enables state-of-the-art blind denoising performance, and near perfect generalization on noise-levels unseen during training. Furthermore, we show that such performance extends to the JDD task and unsupervised learning.