Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2501.09718 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910003454017536 |
|---|---|
| author | Benito, Juan C. Feijoo, Daniel Garcia, Alvaro Conde, Marcos V. |
| author_facet | Benito, Juan C. Feijoo, Daniel Garcia, Alvaro Conde, Marcos V. |
| contents | Low-Light Image Enhancement (LLIE) is a key task in computational photography and imaging. The problem of enhancing images captured during night or in dark environments has been well-studied in the computer vision literature. However, current deep learning-based solutions struggle with efficiency and robustness for real-world scenarios (e.g., scenes with noise, saturated pixels). We propose a lightweight neural network that combines image processing in the frequency and spatial domains. Our baseline method, FLOL, is one of the fastest models for this task, achieving results comparable to the state-of-the-art on popular real-world benchmarks such as LOLv2, LSRW, MIT-5K and UHD-LL. Moreover, we are able to process 1080p images in real-time under 12ms. Code and models at https://github.com/cidautai/FLOL |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_09718 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FLOL: Fast Baselines for Real-World Low-Light Enhancement Benito, Juan C. Feijoo, Daniel Garcia, Alvaro Conde, Marcos V. Computer Vision and Pattern Recognition Robotics Low-Light Image Enhancement (LLIE) is a key task in computational photography and imaging. The problem of enhancing images captured during night or in dark environments has been well-studied in the computer vision literature. However, current deep learning-based solutions struggle with efficiency and robustness for real-world scenarios (e.g., scenes with noise, saturated pixels). We propose a lightweight neural network that combines image processing in the frequency and spatial domains. Our baseline method, FLOL, is one of the fastest models for this task, achieving results comparable to the state-of-the-art on popular real-world benchmarks such as LOLv2, LSRW, MIT-5K and UHD-LL. Moreover, we are able to process 1080p images in real-time under 12ms. Code and models at https://github.com/cidautai/FLOL |
| title | FLOL: Fast Baselines for Real-World Low-Light Enhancement |
| topic | Computer Vision and Pattern Recognition Robotics |
| url | https://arxiv.org/abs/2501.09718 |