SEE is a framework for adaptive brightness adjustment across broad lighting conditions using RGB frames and event-camera data. This repository provides the code for SEE-Net and resources for the SEE-600K dataset.
- Event-guided brightness adjustment: uses RGB frames together with event data for broad-light-range image enhancement.
- Continuous exposure control: supports pixel-level brightness adjustment through an exposure prompt.
- Compact baseline: SEE-Net has approximately 1.9M parameters.
- Large-scale dataset support: built around SEE-600K, which contains 610,126 images with corresponding event data across 202 real-world scenarios.
| Problem Definition | Dataset Samples |
|---|---|
 |
 |
| Input RGB under challenging illumination + event representation -> brightness-adjusted RGB output. | Examples cover low-light, normal-light, high-light, mixed illumination, and event views. |
| Scenario Coverage | Baseline Visualization |
|---|---|
 |
 |
| 202 real-world scenarios with broad scene categories and multiple lighting groups. | SEE-Net uses RGB frames, event data, and a brightness prompt for controllable output exposure. |
The datasets supporting this project are publicly available:
- SEE-600K Dataset
- SDE Dataset: SDE Dataset GitHub
Pretrained checkpoints, evaluation logs, and experiment files for the released baselines are available on Google Drive:
https://drive.google.com/drive/folders/1SWR9YVIrqFEkGGKrv3wSC5RqmMIEYjV2?usp=sharing
The folder is organized by method and includes checkpoints for EIFT_AAAI_SEE, eSl_SEE, EvLight, and SEENet_SEE, along with related logs and TEST_RESULTS.md.
- Tutorial: environment setup, dataset preparation, training, evaluation, inference, and FAQ.
- CodaBench Submission Guide: step-by-step instructions for packaging and submitting results to CodaBench.
- CodaBench Page Description: concise standalone text for the SEE Challenge CodaBench page.
- Test Results: baseline results on the SEE dataset.
- Mini-GT Test Results: baseline results on the SEE mini GT dataset.
Chinese versions are also available:
- Workshop Website: https://eventbasemultimodalvision.github.io
- Competition Page: https://www.codabench.org/competitions/16195/
- Competition Forum: https://www.codabench.org/forums/15944/
- Competition Report: https://arxiv.org/abs/2502.21120
- IMU Registration Tool: https://github.com/yunfanLu/IMU-Registration-Tool
If you use the SEE-600K dataset or SEE-Net in your research, please cite:
@article{lu2025SEE,
title={SEE: See Everything Every Time - Adaptive Brightness Adjustment for Broad Light Range Images via Events},
author={Yunfan Lu, Xiaogang Xu, Hao Lu, Yanlin Qian, Pengteng Li, Huizai Yao, Bin Yang, Junyi Li, Qianyi Cai, Weiyu Guo, Hui Xiong},
year={2025},
}- Yunfan Lu: ylu066@connect.hkust-gz.edu.cn
- GitHub: yunfanLu
