# BlindPSNR: A No-Reference Fidelity Predictor for Low-Light Image Enhancement

arXiv: https://arxiv.org/abs/2607.27628 · HTML: https://mlsanity.com/paper/2607.27628/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 35 references that are in the map.
- First posted on arXiv 30 Jul 2026.

## Builds on (most cited first, 20 of 20)

- [Feature Pyramid Networks for Object Detection](https://mlsanity.com/paper/1612.03144/) (2016)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?](https://mlsanity.com/paper/1703.04977/) (2017)
- [Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields](https://mlsanity.com/paper/2111.12077/) (2021)
- [MUSIQ: Multi-scale Image Quality Transformer](https://mlsanity.com/paper/2108.05997/) (2021)
- [Deep Retinex Decomposition for Low-Light Enhancement](https://mlsanity.com/paper/1808.04560/) (2018)
- [Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement](https://mlsanity.com/paper/2001.06826/) (2020)
- [MANIQA: Multi-dimension Attention Network for No-Reference Image Quality Assessment](https://mlsanity.com/paper/2204.08958/) (2022)
- [TOPIQ: A Top-down Approach from Semantics to Distortions for Image Quality Assessment](https://mlsanity.com/paper/2308.03060/) (2023)
- [Low-Light Image and Video Enhancement Using Deep Learning: A Survey](https://mlsanity.com/paper/2104.10729/) (2021)
- [R2RNet: Low-light Image Enhancement via Real-low to Real-normal Network](https://mlsanity.com/paper/2106.14501/) (2021)
- [HVI: A New Color Space for Low-light Image Enhancement](https://mlsanity.com/paper/2502.20272/) (2025)
- [Learning Conditional Knowledge Distillation for Degraded-Reference Image Quality Assessment](https://mlsanity.com/paper/2108.07948/) (2021)
- [Aleth-NeRF: Illumination Adaptive NeRF with Concealing Field Assumption](https://mlsanity.com/paper/2312.09093/) (2023)
- [FeatEnHancer: Enhancing Hierarchical Features for Object Detection and Beyond Under Low-Light Vision](https://mlsanity.com/paper/2308.03594/) (2023)
- [RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstruction](https://mlsanity.com/paper/2512.23437/) (2025)
- [Content-Variant Reference Image Quality Assessment via Knowledge Distillation](https://mlsanity.com/paper/2202.13123/) (2022)
- [LoLI-Street: Benchmarking Low-Light Image Enhancement and Beyond](https://mlsanity.com/paper/2410.09831/) (2024)
- [Gap-closing Matters: Perceptual Quality Evaluation and Optimization of Low-Light Image Enhancement](https://mlsanity.com/paper/2302.11464/) (2023)
- [Blind Multimodal Quality Assessment of Low-light Images](https://mlsanity.com/paper/2303.10369/) (2023)

## Abstract (arXiv)

Low-light image enhancement (LLIE) methods involve tunable parameters that are typically fixed, often leading to performance degradation when applied across scenes. Manually selecting the best configuration, however, can be time-consuming and not always practical. Peak signal-to-noise ratio (PSNR) is the natural fidelity criterion for automating parameter selection, yet it requires a ground-truth reference that is typically unavailable. To our knowledge, no learning-based method addresses no-reference PSNR prediction for low-light image enhancement; the natural surrogate, no-reference image quality assessment (NR-IQA), targets perceptual quality rather than signal fidelity, and all seven baselines we test achieve 0% top-1 selection accuracy on our benchmark. With paired training data, the ground-truth PSNR is analytically computable, providing exact supervision without a separate teacher network. Building on this, we propose BlindPSNR, a lightweight no-reference network that fuses the enhanced image with the degraded low-light input via windowed cross-attention and estimates PSNR through heteroscedastic regression. While a scalar-regression baseline achieves top-1 accuracy of 54.4%, BlindPSNR raises this to 89.5% with regret dropping from 1.62 dB to 0.026 dB, and generalizes to unseen datasets (SRCC = 0.61-0.67).