DURRNet: Deep Unfolded Single Image Reflection Removal Network

TL;DR

DURRNet is a novel deep unfolded network that combines model-based and learning-based methods to effectively separate reflections from single images, achieving state-of-the-art results.

Contribution

The paper introduces a deep unfolded architecture for reflection removal that integrates interpretability with high performance, using a transform-based prior and deep unrolling techniques.

Findings

Achieves state-of-the-art performance on benchmark datasets.

Effectively separates reflections from images both visually and quantitatively.

Combines model-based optimization with deep learning for interpretability.

Abstract

Single image reflection removal problem aims to divide a reflection-contaminated image into a transmission image and a reflection image. It is a canonical blind source separation problem and is highly ill-posed. In this paper, we present a novel deep architecture called deep unfolded single image reflection removal network (DURRNet) which makes an attempt to combine the best features from model-based and learning-based paradigms and therefore leads to a more interpretable deep architecture. Specifically, we first propose a model-based optimization with transform-based exclusion prior and then design an iterative algorithm with simple closed-form solutions for solving each sub-problems. With the deep unrolling technique, we build the DURRNet with ProxNets to model natural image priors and ProxInvNets which are constructed with invertible networks to impose the exclusion prior. Comprehensive experimental results on commonly used datasets demonstrate that the proposed DURRNet achieves state-of-the-art results both visually and quantitatively.