# Two-Stage Single Image Reflection Removal with Reflection-Aware Guidance

arXiv: https://arxiv.org/abs/2012.00945 · HTML: https://mlsanity.com/paper/2012.00945/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Dec 2020.

## Built on by (most cited first, 22 of 22)

- [Single Image Reflection Separation via Component Synergy](https://mlsanity.com/paper/2308.10027/) (2023)
- [Revisiting Single Image Reflection Removal In the Wild](https://mlsanity.com/paper/2311.17320/) (2023)
- [Reversible Decoupling Network for Single Image Reflection Removal](https://mlsanity.com/paper/2410.08063/) (2024)
- [Removing Reflections from RAW Photos](https://mlsanity.com/paper/2404.14414/) (2024)
- [Flash-Split: 2D Reflection Removal with Flash Cues and Latent Diffusion Separation](https://mlsanity.com/paper/2501.00637/) (2024)
- [A Lightweight Deep Exclusion Unfolding Network for Single Image Reflection Removal](https://mlsanity.com/paper/2503.01938/) (2025)
- [Survey on Single-Image Reflection Removal using Deep Learning Techniques](https://mlsanity.com/paper/2502.08836/) (2025)
- [FIRM: Flexible Interactive Reflection reMoval](https://mlsanity.com/paper/2406.01555/) (2024)
- [OpenRR-1k: A Scalable Dataset for Real-World Reflection Removal](https://mlsanity.com/paper/2506.08299/) (2025)
- [Rectifying Latent Space for Generative Single-Image Reflection Removal](https://mlsanity.com/paper/2512.06358/) (2025)
- [Adaptive Network Combination for Single-Image Reflection Removal: A Domain Generalization Perspective](https://mlsanity.com/paper/2204.01505/) (2022)
- [Flash-Splat: 3D Reflection Removal with Flash Cues and Gaussian Splats](https://mlsanity.com/paper/2410.02764/) (2024)
- [F2T2-HiT: A U-Shaped FFT Transformer and Hierarchical Transformer for Reflection Removal](https://mlsanity.com/paper/2506.05489/) (2025)
- [Single Image Reflection Separation via Dual Prior Interaction Transformer](https://mlsanity.com/paper/2505.12641/) (2025)
- [ReflexSplit: Single Image Reflection Separation via Layer Fusion-Separation](https://mlsanity.com/paper/2601.17468/) (2026)
- [FUMO: Prior-Modulated Diffusion for Single Image Reflection Removal](https://mlsanity.com/paper/2603.19036/) (2026)
- [Reflection Removal Using Recurrent Polarization-to-Polarization Network](https://mlsanity.com/paper/2402.18178/) (2024)
- [SIRR-LMM: Single-image Reflection Removal via Large Multimodal Model](https://mlsanity.com/paper/2601.07209/) (2026)
- [Principled Reflection Separation via Nonlinear Superposition and Feature Interaction](https://mlsanity.com/paper/2606.02831/) (2026)
- [Reflection Separation from a Single Image via Joint Latent Diffusion](https://mlsanity.com/paper/2606.04107/) (2026)
- [ReLo-IRR: Reflection-Guided LoRA Framework for Image Reflection Removal](https://mlsanity.com/paper/2607.02957/) (2026)
- [LowAux-RDNet: Low-Pass Residual Supervision with Scene-Balanced Real-World Training for Single-Image Reflection Removal](https://mlsanity.com/paper/2607.22707/) (2026)

## Abstract (arXiv)

Removing undesired reflection from an image captured through a glass surface is a very challenging problem with many practical application scenarios. For improving reflection removal, cascaded deep models have been usually adopted to estimate the transmission in a progressive manner. However, most existing methods are still limited in exploiting the result in prior stage for guiding transmission estimation. In this paper, we present a novel two-stage network with reflection-aware guidance (RAGNet) for single image reflection removal (SIRR). To be specific, the reflection layer is firstly estimated due to that it generally is much simpler and is relatively easier to estimate. Reflectionaware guidance (RAG) module is then elaborated for better exploiting the estimated reflection in predicting transmission layer. By incorporating feature maps from the estimated reflection and observation, RAG can be used (i) to mitigate the effect of reflection from the observation, and (ii) to generate mask in partial convolution for mitigating the effect of deviating from linear combination hypothesis. A dedicated mask loss is further presented for reconciling the contributions of encoder and decoder features. Experiments on five commonly used datasets demonstrate the quantitative and qualitative superiority of our RAGNet in comparison to the state-of-the-art SIRR methods. The source code and pre-trained model are available at https://github.com/liyucs/RAGNet.