Single-image reflection removal via self-supervised diffusion models

TL;DR

This paper introduces a self-supervised diffusion model approach for single-image reflection removal that does not require paired training data, achieving superior results on multiple datasets.

Contribution

It proposes a novel hybrid method combining cycle-consistency with diffusion probabilistic models for reflection removal without paired data.

Findings

Outperforms state-of-the-art methods on SIR$^2$, FRR, and MRR datasets.

Effectively models decomposition and synthesis of reflection and transmission images.

Demonstrates robustness and high-quality results in reflection removal.

Abstract

Reflections often degrade the visual quality of images captured through transparent surfaces, and reflection removal methods suffers from the shortage of paired real-world samples.This paper proposes a hybrid approach that combines cycle-consistency with denoising diffusion probabilistic models (DDPM) to effectively remove reflections from single images without requiring paired training data. The method introduces a Reflective Removal Network (RRN) that leverages DDPMs to model the decomposition process and recover the transmission image, and a Reflective Synthesis Network (RSN) that re-synthesizes the input image using the separated components through a nonlinear attention-based mechanism. Experimental results demonstrate the effectiveness of the proposed method on the SIR$^2$, Flash-Based Reflection Removal (FRR) Dataset, and a newly introduced Museum Reflection Removal (MRR) dataset, showing superior performance compared to state-of-the-art methods.