mlsanityData to 14 Oct 2026

Removing Reflections from RAW Photos

arXiv 2404.14414 ↗First posted 28 Feb 2024cs.CVcited by 10

Not on the map. Why?

Abstract (arXiv)

We describe a system to remove real-world reflections from images for consumer photography. Our system operates on linear (RAW) photos, and accepts an optional contextual photo looking in the opposite direction (e.g., the "selfie" camera on a mobile device). This optional photo disambiguates what should be considered the reflection. The system is trained solely on synthetic mixtures of real RAW photos, which we combine using a reflection simulation that is photometrically and geometrically accurate. Our system comprises a base model that accepts the captured photo and optional context photo as input, and runs at 256p, followed by an up-sampling model that transforms 256p images to full resolution. The system produces preview images at 1K in 4.5-6.5s on a MacBook or iPhone 14 Pro. We show SOTA results on RAW photos that were captured in the field to embody typical consumer photos, and show that training on RAW simulation data improves performance more than the architectural variations among prior works.

Key facts. Cited by 10 papers in the map (data through 14 Oct 2026). Builds on 26 of its 59 references that are in the map. First posted on arXiv 28 Feb 2024.

Builds on

26 of 59 references.

References · 26

  1. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,364
  2. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,442
  3. 2020Training Generative Adversarial Networks with Limited Datacited by 1,186
  4. 2019EfficientDet: Scalable and Efficient Object Detectioncited by 1,020
  5. 2021Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Datacited by 970
  6. 2024Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wildcited by 286
  7. 2018Unprocessing Images for Learned Raw Denoisingcited by 250
  8. 2017Learning to Predict Indoor Illumination from a Single Imagecited by 161
  9. 2018Single Image Reflection Separation with Perceptual Lossescited by 98
  10. 2020Polarized Reflection Removal with Perfect Alignment in the Wildcited by 69
  11. 2017A Generic Deep Architecture for Single Image Reflection Removal and Image Smoothingcited by 60
  12. 2019Single Image Reflection Removal through Cascaded Refinementcited by 58
14 more
  1. 2019Single Image Reflection Removal Exploiting Misaligned Training Data and Network Enhancementscited by 54
  2. 2017A Perceptual Measure for Deep Single Image Camera Calibrationcited by 45
  3. 2020Learning to See Through Obstructionscited by 38
  4. 2023Single Image Reflection Separation via Component Synergycited by 38
  5. 2021Robust Reflection Removal with Reflection-free Flash-only Cuescited by 34
  6. 2023Revisiting Single Image Reflection Removal In the Wildcited by 33
  7. 2020Cross-Camera Convolutional Color Constancycited by 30
  8. 2020CIE XYZ Net: Unprocessing Images for Low-Level Computer Vision Taskscited by 29
  9. 2018The Visual Centrifuge: Model-Free Layered Video Representationscited by 23
  10. 2020Two-Stage Single Image Reflection Removal with Reflection-Aware Guidancecited by 22
  11. 2017Separating Reflection and Transmission Images in the Wildcited by 19
  12. 2019Fast Single Image Reflection Suppression via Convex Optimizationcited by 17
  13. 2021A Categorized Reflection Removal Dataset with Diverse Real-world Scenescited by 12
  14. 2015Automatic Layer Separation using Light Field Imagingcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 2; 2025: 1; 2026: 7.

10 papers.

  1. 2024Flash-Split: 2D Reflection Removal with Flash Cues and Latent Diffusion Separationcited by 10
  2. 2024Flash-Splat: 3D Reflection Removal with Flash Cues and Gaussian Splatscited by 3
  3. 2026RAWild: Sensor-Agnostic RAW Object Detection via Physics-Guided Curve and Grid Modelingcited by 3
  4. 2025Raw-JPEG Adapter: Efficient Raw Image Compression with JPEGcited by 0
  5. 2026SIRR-LMM: Single-image Reflection Removal via Large Multimodal Modelcited by 0
  6. 2026MERIT: Multi-domain Efficient RAW Image Translationcited by 0
  7. 2026Linear Image Generation by Synthesizing Exposure Bracketscited by 0
  8. 2026Principled Reflection Separation via Nonlinear Superposition and Feature Interactioncited by 0
  9. 2026PRISM: Latent Composition Consistency for Single-Image Reflection Removalcited by 0
  10. 2026OccludeBench: Can VLMs Find What's Missing?cited by 0