Learning to Remove Lens Flare in Event Camera
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912756099186688 |
|---|---|
| author | Han, Haiqian Kong, Lingdong Li, Jianing Liang, Ao Zhu, Chengtao Lyu, Jiacheng Ng, Lai Xing Ji, Xiangyang Ooi, Wei Tsang Cottereau, Benoit R. |
| author_facet | Han, Haiqian Kong, Lingdong Li, Jianing Liang, Ao Zhu, Chengtao Lyu, Jiacheng Ng, Lai Xing Ji, Xiangyang Ooi, Wei Tsang Cottereau, Benoit R. |
| contents | Event cameras have the potential to revolutionize vision systems with their high temporal resolution and dynamic range, yet they remain susceptible to lens flare, a fundamental optical artifact that causes severe degradation. In event streams, this optical artifact forms a complex, spatio-temporal distortion that has been largely overlooked. We present E-Deflare, the first systematic framework for removing lens flare from event camera data. We first establish the theoretical foundation by deriving a physics-grounded forward model of the non-linear suppression mechanism. This insight enables the creation of the E-Deflare Benchmark, a comprehensive resource featuring a large-scale simulated training set, E-Flare-2.7K, and the first-ever paired real-world test set, E-Flare-R, captured by our novel optical system. Empowered by this benchmark, we design E-DeflareNet, which achieves state-of-the-art restoration performance. Extensive experiments validate our approach and demonstrate clear benefits for downstream tasks. Code and datasets are publicly available. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_09016 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Learning to Remove Lens Flare in Event Camera Han, Haiqian Kong, Lingdong Li, Jianing Liang, Ao Zhu, Chengtao Lyu, Jiacheng Ng, Lai Xing Ji, Xiangyang Ooi, Wei Tsang Cottereau, Benoit R. Computer Vision and Pattern Recognition Event cameras have the potential to revolutionize vision systems with their high temporal resolution and dynamic range, yet they remain susceptible to lens flare, a fundamental optical artifact that causes severe degradation. In event streams, this optical artifact forms a complex, spatio-temporal distortion that has been largely overlooked. We present E-Deflare, the first systematic framework for removing lens flare from event camera data. We first establish the theoretical foundation by deriving a physics-grounded forward model of the non-linear suppression mechanism. This insight enables the creation of the E-Deflare Benchmark, a comprehensive resource featuring a large-scale simulated training set, E-Flare-2.7K, and the first-ever paired real-world test set, E-Flare-R, captured by our novel optical system. Empowered by this benchmark, we design E-DeflareNet, which achieves state-of-the-art restoration performance. Extensive experiments validate our approach and demonstrate clear benefits for downstream tasks. Code and datasets are publicly available. |
| title | Learning to Remove Lens Flare in Event Camera |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.09016 |