# Using KL-Divergence to Focus Frequency Information in Low-Light Image Enhancement

arXiv: https://arxiv.org/abs/2509.13083 · HTML: https://mlsanity.com/paper/2509.13083/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 49 references that are in the map.
- First posted on arXiv 16 Sep 2025.

## Builds on (most cited first, 26 of 26)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [Restormer: Efficient Transformer for High-Resolution Image Restoration](https://mlsanity.com/paper/2111.09881/) (2021)
- [Deep Retinex Decomposition for Low-Light Enhancement](https://mlsanity.com/paper/1808.04560/) (2018)
- [EnlightenGAN: Deep Light Enhancement without Paired Supervision](https://mlsanity.com/paper/1906.06972/) (2019)
- [Tutorial on Variational Autoencoders](https://mlsanity.com/paper/1606.05908/) (2016)
- [Kindling the Darkness: A Practical Low-light Image Enhancer](https://mlsanity.com/paper/1905.04161/) (2019)
- [Learning Enriched Features for Real Image Restoration and Enhancement](https://mlsanity.com/paper/2003.06792/) (2020)
- [A Fourier-based Framework for Domain Generalization](https://mlsanity.com/paper/2105.11120/) (2021)
- [Low-Light Image Enhancement with Normalizing Flow](https://mlsanity.com/paper/2109.05923/) (2021)
- [Ultra-High-Definition Low-Light Image Enhancement: A Benchmark and Transformer-Based Method](https://mlsanity.com/paper/2212.11548/) (2022)
- [Rethinking Rotated Object Detection with Gaussian Wasserstein Distance Loss](https://mlsanity.com/paper/2101.11952/) (2021)
- [Diff-Retinex: Rethinking Low-light Image Enhancement with A Generative Diffusion Model](https://mlsanity.com/paper/2308.13164/) (2023)
- [R2RNet: Low-light Image Enhancement via Real-low to Real-normal Network](https://mlsanity.com/paper/2106.14501/) (2021)
- [Fourier Space Losses for Efficient Perceptual Image Super-Resolution](https://mlsanity.com/paper/2106.00783/) (2021)
- [Low-light Image Enhancement via Breaking Down the Darkness](https://mlsanity.com/paper/2111.15557/) (2021)
- [FourLLIE: Boosting Low-Light Image Enhancement by Fourier Frequency Information](https://mlsanity.com/paper/2308.03033/) (2023)
- [HVI: A New Color Space for Low-light Image Enhancement](https://mlsanity.com/paper/2502.20272/) (2025)
- [Retinexmamba: Retinex-based Mamba for Low-light Image Enhancement](https://mlsanity.com/paper/2405.03349/) (2024)
- [Zero-Reference Low-Light Enhancement via Physical Quadruple Priors](https://mlsanity.com/paper/2403.12933/) (2024)
- [Tutorial: Deriving the Standard Variational Autoencoder (VAE) Loss Function](https://mlsanity.com/paper/1907.08956/) (2019)
- [LYT-NET: Lightweight YUV Transformer-based Network for Low-light Image Enhancement](https://mlsanity.com/paper/2401.15204/) (2024)
- [Generalizable Medical Image Segmentation via Random Amplitude Mixup and Domain-Specific Image Restoration](https://mlsanity.com/paper/2208.03901/) (2022)
- [DMFourLLIE: Dual-Stage and Multi-Branch Fourier Network for Low-Light Image Enhancement](https://mlsanity.com/paper/2412.00683/) (2024)
- [YOLC: You Only Look Clusters for Tiny Object Detection in Aerial Images](https://mlsanity.com/paper/2404.06180/) (2024)
- [Low-Light Image Enhancement by Learning Contrastive Representations in Spatial and Frequency Domains](https://mlsanity.com/paper/2303.13412/) (2023)

## Abstract (arXiv)

In the Fourier domain, luminance information is primarily encoded in the amplitude spectrum, while spatial structures are captured in the phase components. The traditional Fourier Frequency information fitting employs pixel-wise loss functions, which tend to focus excessively on local information and may lead to global information loss. In this paper, we present LLFDisc, a U-shaped deep enhancement network that integrates cross-attention and gating mechanisms tailored for frequency-aware enhancement. We propose a novel distribution-aware loss that directly fits the Fourier-domain information and minimizes their divergence using a closed-form KL-Divergence objective. This enables the model to align Fourier-domain information more robustly than with conventional MSE-based losses. Furthermore, we enhance the perceptual loss based on VGG by embedding KL-Divergence on extracted deep features, enabling better structural fidelity. Extensive experiments across multiple benchmarks demonstrate that LLFDisc achieves state-of-the-art performance in both qualitative and quantitative evaluations. Our code will be released at: https://github.com/YanXY000/LLFDisc