Rethinking RGB Color Representation for Image Restoration Models
Not on the map. Why?
Abstract (arXiv)
Image restoration models are typically trained with a pixel-wise distance loss defined over the RGB color representation space, which is well known to be a source of blurry and unrealistic textures in the restored images. The reason, we believe, is that the three-channel RGB space is insufficient for supervising the restoration models. To this end, we augment the representation to hold structural information of local neighborhoods at each pixel while keeping the color information and pixel-grainedness unharmed. The result is a new representation space, dubbed augmented RGB (aRGB) space. Substituting the underlying representation space for the per-pixel losses facilitates the training of image restoration models, thereby improving the performance without affecting the evaluation phase. Notably, when combined with auxiliary objectives such as adversarial or perceptual losses, our aRGB space consistently improves overall metrics by reconstructing both color and local structures, overcoming the conventional perception-distortion trade-off.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 36 of its 56 references that are in the map. First posted on arXiv 5 Feb 2024.
Builds on
36 of 56 references.
References · 36
- 2017Attention Is All You Needcited by 59,932
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
- 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
- 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
- 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
- 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
- 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
- 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,413
- 2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 3,014
- 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,678
- 2016Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Networkcited by 1,898
- 2021SwinIR: Image Restoration Using Swin Transformercited by 1,869
24 more
- 2021Restormer: Efficient Transformer for High-Resolution Image Restorationcited by 1,618
- 2017Enhanced Deep Residual Networks for Single Image Super-Resolutioncited by 1,543
- 2016Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoisingcited by 1,352
- 2018On the Spectral Bias of Neural Networkscited by 1,321
- 2018Image Super-Resolution Using Very Deep Residual Channel Attention Networkscited by 1,256
- 2014Image Super-Resolution Using Deep Convolutional Networkscited by 1,218
- 2017Deep Image Priorcited by 1,097
- 2015Accurate Image Super-Resolution Using Very Deep Convolutional Networkscited by 902
- 2021Multi-Stage Progressive Image Restorationcited by 813
- 2020Image Quality Assessment: Unifying Structure and Texture Similaritycited by 792
- 2016Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurringcited by 722
- 2022Simple Baselines for Image Restorationcited by 701
- 2022Mixture-of-Experts with Expert Choice Routingcited by 611
- 2021Designing a Practical Degradation Model for Deep Blind Image Super-Resolutioncited by 524
- 2017Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolutioncited by 452
- 2018Recovering Realistic Texture in Image Super-resolution by Deep Spatial Feature Transformcited by 369
- 2015Sketch-based Manga Retrieval using Manga109 Datasetcited by 326
- 2021Rethinking Coarse-to-Fine Approach in Single Image Deblurringcited by 324
- 2015Deeply-Recursive Convolutional Network for Image Super-Resolutioncited by 317
- 2020Multi-Scale Progressive Fusion Network for Single Image Derainingcited by 259
- 2020Human-Aware Motion Deblurringcited by 174
- 2019RankSRGAN: Generative Adversarial Networks with Ranker for Image Super-Resolutioncited by 145
- 2021FastMoE: A Fast Mixture-of-Expert Training Systemcited by 110
- 2021Improving Image Restoration by Revisiting Global Information Aggregationcited by 60
Built on by
Citing papers per year: 2025: 2.
2 papers.