Research map / Image generation / Image restoration
BiEvLight: Bi-level Learning of Task-Aware Event Refinement for Low-Light Image Enhancement
Zishu Yao, Xiang-Xiang Su, Shengning Zhou, Guang-Yong Chen, Guodong Fan, Xing Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add event denoising to the enhancement network (ablation) | BiEvLight base · SDE-in | PSNR*: 21.9412 -> 26.0023 | Improved |
| add event denoising to the enhancement network (ablation) | BiEvLight base · SDE-out | PSNR: 21.8932 -> 24.3599 | Improved |
| guide denoising with gradient of reflectance from low-light image only (vs no denoising) (ablation) | BiEvLight base · SDE-in | PSNR*: 21.9412 -> 24.6423 | Improved |
| use the spatially-adaptive gradient-guided denoising strategy instead of low-light-only gradient (ablation) | BiEvLight base · SDE-in | PSNR*: 24.6423 -> 25.5993 | Improved |
| apply bilevel learning instead of joint learning (ablation) | BiEvLight · SDE-in | PSNR: 21.9812 -> 22.8680 | Improved |
| apply bilevel learning instead of joint learning (ablation) | BiEvLight · SDE-out | PSNR*: 24.5623 -> 26.1617 | Improved |
| apply bilevel learning instead of alternating learning (ablation) | BiEvLight · SDE-in | PSNR*: 25.3691 -> 26.0023 | Improved |
| apply bilevel learning instead of alternating learning (ablation) | BiEvLight · SDE-out | PSNR: 24.1234 -> 24.3599 | Improved |
| BiEvLight vs prior best EvLight | SDE-in · SDE-in | PSNR: 22.1880 -> 22.8680 | Improved |
| BiEvLight vs prior best EvLight | SDE-out · SDE-out | PSNR: 22.4372 -> 24.3599 | Improved |
| BiEvLight vs prior best EvLight | SDSD-in · SDSD-in | PSNR: 29.3563 -> 30.7576 | Improved |
| BiEvLight vs prior best EvLight on PSNR* | SDSD-out · SDSD-out | PSNR*: 30.3066 -> 30.4579 | Improved |
Closest papers by meaning
- Towards Robust Event-guided Low-Light Image Enhancement: A Large-Scale Real-World Event-Image Dataset and Novel Approach (CVPR 2024 · similarity 0.88)
- LED: A Large-scale Real-world Paired Dataset for Event Camera Denoising (CVPR 2024 · similarity 0.88)
- Coherent Event Guided Low-Light Video Enhancement (ICCV 2023 · similarity 0.88)
- EDformer: Transformer-Based Event Denoising Across Varied Noise Levels (ECCV 2024 · similarity 0.87)
- Event-Guided Consistent Video Enhancement with Modality-Adaptive Diffusion Pipeline (NeurIPS 2025 · similarity 0.87)
- Interpretable Unsupervised Joint Denoising and Enhancement for Real-World low-light Scenarios (ICLR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
