Research map / 3D vision / NeRF and radiance fields
High-speed Imaging through Turbulence with Event-based Light Fields
Yu-Hsiang Huang, Levi Burner, Sachin Shah, Ziyuan Qu, Adithya Pediredla, Christopher Metzler
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 |
|---|---|---|---|
| use 3x3 light-field views instead of single view | E2VID · REDS simulated | LPIPS: 0.4607 -> 0.4332 | Improved |
| use 3x3 light field instead of single view at matched pixel count (768x768) | E2VID · REDS simulated | ITF: 22.71 -> 25.16 | Improved |
| use light field instead of single view | E2VID · Tabletop | E_warp: 0.0085 -> 0.0051 | Improved |
| use light field instead of single view | E2VID · Tabletop | PSNR: 12.37 -> 12.99 | Improved |
| use 5x5 views instead of 3x3 (ablation) | E2VID · REDS simulated | ITF: 25.16 -> 29.85 | Improved |
| use 5x5 views instead of 3x3 (ablation) | E2VID · REDS simulated | E_warp: 0.0022 -> 0.0010 | Improved |
| use more views with fewer pixels each (5x5 vs 3x3) (ablation) | E2VID · REDS simulated | LPIPS: 0.433 -> 0.443 | Hurt |
| replace LPIPS loss with MSE (ablation) | E2VID light field · REDS simulated | PSNR / LPIPS: PSNR 15.40 -> 16.08; LPIPS 0.433 -> 0.579 | Mixed |
| remove spectral normalization on ConvLSTM (ablation) | E2VID light field · REDS simulated | qualitative | Improved |
| postprocess E2VID+ output with MambaTM (ablation) | E2VID+ · REDS simulated | LPIPS: 0.5854 -> 0.5602 | Improved |
| postprocess FireNet+ output with MambaTM (ablation) | FireNet+ · REDS simulated | PSNR: 12.84 -> 12.68 | Hurt |
| use end-to-end light-field training instead of existing E2V methods retrained (ablation) | E2VID · REDS simulated | PSNR: 14.51 -> 15.40 | Improved |
Closest papers by meaning
- Event Fields: Capturing Light Fields at High Speed, Resolution, and Dynamic Range (CVPR 2025 · similarity 0.92)
- EventSplat: 3D Gaussian Splatting from Moving Event Cameras for Real-time Rendering (CVPR 2025 · similarity 0.86)
- Event-3DGS: Event-based 3D Reconstruction Using 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.86)
- Time-Efficient Light-Field Acquisition Using Coded Aperture and Events (CVPR 2024 · similarity 0.86)
- Generalized Event Cameras (CVPR 2024 · similarity 0.85)
- EF-3DGS: Event-Aided Free-Trajectory 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.85)
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.
