Research map / Video / Video segmentation and tracking
Deblur e-NeRF: NeRF from Motion-Blurred Events under High-speed or Low-light Conditions
Weng Fei Low, Gim Hee Lee
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 the physical pixel bandwidth model to event-based NeRF training (ablation) | Deblur e-NeRF · Synthetic hard (v=4x, 10 lux) | PSNR: 18.31 -> 26.15 | Improved |
| Add the pixel bandwidth model under mild blur (ablation) | Deblur e-NeRF · Synthetic easy (v=0.125x, 100000 lux) | PSNR: 28.75 -> 29.00 | No clear effect |
| Jointly optimize contrast thresholds and refractory period in Robust e-NeRF (ablation) | Robust e-NeRF · Synthetic hard (v=4x, 10 lux) | PSNR: 18.42 -> 18.51 | No clear effect |
| Replace target-normalized gradient loss with threshold-normalized total variation loss (no blur) (ablation) | Robust e-NeRF · Blur-free synthetic upper bound | PSNR: 28.48 -> 29.43 | Improved |
| Jointly optimize pixel bandwidth model parameters, initialized at 4x true value (ablation) | Deblur e-NeRF · Synthetic hard | PSNR: 26.15 -> 25.59 | Mixed |
| Increase training batch size from 1/8x to 1x of baselines (ablation) | Deblur e-NeRF · Synthetic lego hard | PSNR: 27.69 -> 30.72 | Improved |
| Increase importance-sampling input sample size from 1 to 30 (ablation) | Deblur e-NeRF · Synthetic lego hard | PSNR: 18.46 -> 27.69 | Improved |
| Increase input sample size from 30 to 75 (ablation) | Deblur e-NeRF · Synthetic lego hard | PSNR: 27.69 -> 28.21 | No clear effect |
| Add image deblur or blur-aware models (MPRNet, Deblur-NeRF) to E2VID + NeRF | E2VID + NeRF · Synthetic and real | PSNR | No clear effect |
| Use Deblur e-NeRF instead of Robust e-NeRF under low illuminance | Event NeRF reconstruction · Synthetic 10 lux | PSNR: 22.72 -> 28.62 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Robust e-NeRF — Robust e-NeRF: NeRF from Sparse & Noisy Events under Non-Uniform Motion (ICCV 2023) · better on 10 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Mitigating Motion Blur in Neural Radiance Fields with Events and Frames (CVPR 2024 · similarity 0.92)
- AE-NeRF: Augmenting Event-Based Neural Radiance Fields for Non-ideal Conditions and Larger Scenes (AAAI 2025 · similarity 0.90)
- DynaMoDe-NeRF: Motion-aware Deblurring Neural Radiance Field for Dynamic Scenes (CVPR 2025 · similarity 0.89)
- Frequency-aware Event-based Video Deblurring for Real-World Motion Blur (CVPR 2024 · similarity 0.88)
- Latency Correction for Event-guided Deblurring and Frame Interpolation (CVPR 2024 · similarity 0.88)
- LED: A Large-scale Real-world Paired Dataset for Event Camera Denoising (CVPR 2024 · similarity 0.88)
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.
