# Learning a self-supervised tone mapping operator via feature contrast masking loss

arXiv: https://arxiv.org/abs/2110.09866 · HTML: https://mlsanity.com/paper/2110.09866/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 84 references that are in the map.
- First posted on arXiv 19 Oct 2021.

## Builds on (most cited first, 14 of 14)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://mlsanity.com/paper/1704.04861/) (2017)
- [Perceptual Losses for Real-Time Style Transfer and Super-Resolution](https://mlsanity.com/paper/1603.08155/) (2016)
- [HDR image reconstruction from a single exposure using deep CNNs](https://mlsanity.com/paper/1710.07480/) (2017)
- [Learning to Predict Indoor Illumination from a Single Image](https://mlsanity.com/paper/1704.00090/) (2017)
- [Deep High Dynamic Range Imaging with Large Foreground Motions](https://mlsanity.com/paper/1711.08937/) (2017)
- [Single Image HDR Reconstruction Using a CNN with Masked Features and Perceptual Loss](https://mlsanity.com/paper/2005.07335/) (2020)
- [Image Correction via Deep Reciprocating HDR Transformation](https://mlsanity.com/paper/1804.04371/) (2018)
- [Deep Tone Mapping Operator for High Dynamic Range Images](https://mlsanity.com/paper/1908.04197/) (2019)
- [Enhancement of Underwater Images with Statistical Model of Background Light and Optimization of Transmission Map](https://mlsanity.com/paper/1906.08673/) (2019)
- [LAPRAN: A Scalable Laplacian Pyramid Reconstructive Adversarial Network for Flexible Compressive Sensing Reconstruction](https://mlsanity.com/paper/1807.09388/) (2018)
- [Deep Feature Consistent Deep Image Transformations: Downscaling, Decolorization and HDR Tone Mapping](https://mlsanity.com/paper/1707.09482/) (2017)
- [Explorable Tone Mapping Operators](https://mlsanity.com/paper/2010.10000/) (2020)

## Built on by (most cited first, 6 of 6)

- [GlowGAN: Unsupervised Learning of HDR Images from LDR Images in the Wild](https://mlsanity.com/paper/2211.12352/) (2022)
- [Cinematic Gaussians: Real-Time HDR Radiance Fields with Depth of Field](https://mlsanity.com/paper/2406.07329/) (2024)
- [HDRGS: High Dynamic Range Gaussian Splatting](https://mlsanity.com/paper/2408.06543/) (2024)
- [Perceptual Tone Mapping Model for High Dynamic Range Imaging](https://mlsanity.com/paper/2309.16975/) (2023)
- [Style-Aware Gloss Control for Generative Non-Photorealistic Rendering](https://mlsanity.com/paper/2602.16611/) (2026)
- [DRIFT: Deep Restoration, ISP Fusion, and Tone-mapping](https://mlsanity.com/paper/2604.03402/) (2026)

## Abstract (arXiv)

High Dynamic Range (HDR) content is becoming ubiquitous due to the rapid development of capture technologies. Nevertheless, the dynamic range of common display devices is still limited, therefore tone mapping (TM) remains a key challenge for image visualization. Recent work has demonstrated that neural networks can achieve remarkable performance in this task when compared to traditional methods, however, the quality of the results of these learning-based methods is limited by the training data. Most existing works use as training set a curated selection of best-performing results from existing traditional tone mapping operators (often guided by a quality metric), therefore, the quality of newly generated results is fundamentally limited by the performance of such operators. This quality might be even further limited by the pool of HDR content that is used for training. In this work we propose a learning-based self-supervised tone mapping operator that is trained at test time specifically for each HDR image and does not need any data labeling. The key novelty of our approach is a carefully designed loss function built upon fundamental knowledge on contrast perception that allows for directly comparing the content in the HDR and tone mapped images. We achieve this goal by reformulating classic VGG feature maps into feature contrast maps that normalize local feature differences by their average magnitude in a local neighborhood, allowing our loss to account for contrast masking effects. We perform extensive ablation studies and exploration of parameters and demonstrate that our solution outperforms existing approaches with a single set of fixed parameters, as confirmed by both objective and subjective metrics.