# Semi-LAR: Semi-supervised Contrastive Learning with Linear Attention for Removal of Nighttime Flares

arXiv: https://arxiv.org/abs/2605.18156 · HTML: https://mlsanity.com/paper/2605.18156/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 52 references that are in the map.
- First posted on arXiv 18 May 2026.

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

- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Rich feature hierarchies for accurate object detection and semantic segmentation](https://mlsanity.com/paper/1311.2524/) (2013)
- [Restormer: Efficient Transformer for High-Resolution Image Restoration](https://mlsanity.com/paper/2111.09881/) (2021)
- [Self-training with Noisy Student improves ImageNet classification](https://mlsanity.com/paper/1911.04252/) (2019)
- [MUSIQ: Multi-scale Image Quality Transformer](https://mlsanity.com/paper/2108.05997/) (2021)
- [Uformer: A General U-Shaped Transformer for Image Restoration](https://mlsanity.com/paper/2106.03106/) (2021)
- [Holistically-Nested Edge Detection](https://mlsanity.com/paper/1504.06375/) (2015)
- [Contrastive Learning for Compact Single Image Dehazing](https://mlsanity.com/paper/2104.09367/) (2021)
- [Single Image Reflection Separation with Perceptual Losses](https://mlsanity.com/paper/1806.05376/) (2018)
- [How to Train Neural Networks for Flare Removal](https://mlsanity.com/paper/2011.12485/) (2020)
- [Flare7K: A Phenomenological Nighttime Flare Removal Dataset](https://mlsanity.com/paper/2210.06570/) (2022)
- [Enhancing Visibility in Nighttime Haze Images Using Guided APSF and Gradient Adaptive Convolution](https://mlsanity.com/paper/2308.01738/) (2023)
- [Flare7K++: Mixing Synthetic and Real Datasets for Nighttime Flare Removal and Beyond](https://mlsanity.com/paper/2306.04236/) (2023)
- [A Semi-supervised Nighttime Dehazing Baseline with Spatial-Frequency Aware and Realistic Brightness Constraint](https://mlsanity.com/paper/2403.18548/) (2024)
- [Improving Lens Flare Removal with General Purpose Pipeline and Multiple Light Sources Recovery](https://mlsanity.com/paper/2308.16460/) (2023)
- [Automatic Flare Spot Artifact Detection and Removal in Photographs](https://mlsanity.com/paper/2103.04384/) (2021)
- [MFDNet: Multi-Frequency Deflare Network for Efficient Nighttime Flare Removal](https://mlsanity.com/paper/2406.18079/) (2024)
- [MIPI 2023 Challenge on Nighttime Flare Removal: Methods and Results](https://mlsanity.com/paper/2305.13770/) (2023)
- [Disentangle Nighttime Lens Flares: Self-supervised Generation-based Lens Flare Removal](https://mlsanity.com/paper/2502.10714/) (2025)
- [FlareX: A Physics-Informed Dataset for Lens Flare Removal via 2D Synthesis and 3D Rendering](https://mlsanity.com/paper/2510.09995/) (2025)
- [Toward Real Flare Removal: A Comprehensive Pipeline and A New Benchmark](https://mlsanity.com/paper/2306.15884/) (2023)
- [DFDNet: Dynamic Frequency-Guided De-Flare Network](https://mlsanity.com/paper/2507.17489/) (2025)
- [DeflareMamba: Hierarchical Vision Mamba for Contextually Consistent Lens Flare Removal](https://mlsanity.com/paper/2508.02113/) (2025)
- [SLCFormer: Spectral-Local Context Transformer with Physics-Grounded Flare Synthesis for Nighttime Flare Removal](https://mlsanity.com/paper/2512.15221/) (2025)
- [Semi-LLIE: Semi-supervised Contrastive Learning with Mamba-based Low-light Image Enhancement](https://mlsanity.com/paper/2409.16604/) (2024)

## Abstract (arXiv)

Lens flare removal is challenging due to the large spatial extent of flare artifacts and their entanglement with scene structures, while existing methods heavily rely on large-scale paired data. We propose a semi-supervised flare removal framework that enables stable learning from unlabeled images by jointly addressing pseudo-label reliability and representation discrimination. We propose an adaptive pseudo-label repository that progressively refines pseudo supervision through no-reference quality assessment, momentum-based updates, and invalid label filtering, effectively mitigating error accumulation. Moreover, we propose a flare-aware contrastive loss that explicitly treats flare-contaminated inputs as negatives and performs patch-level contrastive learning, encouraging representations that are discriminative against flare patterns while remaining consistent with reliable pseudo targets. Extensive experiments on multiple flare benchmarks demonstrate that the proposed framework is model-agnostic and consistently improves performance and robustness.