Research map / 3D vision / Depth and stereo
Learning to See Through With Events
Lei Yu, Xiang Zhang, Wei Liao, Wen Tao Yang, Gui-Song Xia
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 |
|---|---|---|---|
| replace CNN encoder with SNN encoder (ablation) | E-SAI hybrid network · SAI dataset (indoor) | PSNR: 26.53 -> 29.55 | Improved |
| replace CNN encoder with SNN encoder (ablation) | E-SAI hybrid network · SAI dataset (outdoor) | PSNR: 15.98 -> 20.39 | Improved |
| use SNN encoder instead of ConvLSTM RNN encoder (ablation) | E-SAI hybrid network · SAI dataset (indoor) | PSNR: 28.38 -> 29.55 | Improved |
| use ConvLSTM RNN encoder instead of CNN encoder (ablation) | E-SAI+CNN · SAI dataset (outdoor) | PSNR: 15.98 -> 18.26 | Improved |
| use CNN decoder instead of 5-layer SNN decoder (ablation) | E-SAI hybrid network · SAI dataset (indoor) | PSNR: 23.34 -> 29.55 | Improved |
| use CNN decoder instead of 7-layer SNN decoder (ablation) | E-SAI hybrid network · SAI dataset (indoor) | PSNR: 7.33 -> 29.55 | Improved |
| increase time intervals N from 1 to 30 (ablation) | E-SAI+Hybrid · SAI dataset (indoor) | PSNR: 29.30 -> 30.71 | Improved |
| increase time intervals N from 1 to 30 (ablation) | E-SAI+Hybrid · SAI dataset (outdoor) | PSNR: 17.80 -> 20.39 | Improved |
| use auto refocus (Refocus-Net) instead of manual refocus | E-SAI+Hybrid · SAI dataset (indoor) | PSNR: 30.71 -> 29.55 | Hurt |
| use auto refocus (Refocus-Net) instead of manual refocus | E-SAI+Hybrid · SAI dataset (outdoor) | PSNR: 20.39 -> 20.39 | No clear effect |
| event-based SAI (hybrid) instead of F-SAI+CNN | E-SAI+Hybrid (M) · SAI dataset (indoor) | PSNR: 25.01 -> 30.71 | Improved |
| event-based SAI (hybrid) instead of F-SAI+CNN | E-SAI+Hybrid (M) · SAI dataset (outdoor) | LPIPS: 0.1347 -> 0.0981 | Improved |
Closest papers by meaning
- Event-Based Synthetic Aperture Imaging With a Hybrid Network (CVPR 2021 · similarity 0.96)
- Synthetic Aperture Imaging With Events and Frames (CVPR 2022 · similarity 0.90)
- REDIR: Refocus-free Event-based De-occlusion Image Reconstruction (ECCV 2024 · similarity 0.88)
- EvFocus: Learning to Reconstruct Sharp Images from Out-of-Focus Event Streams (ICML 2025 · similarity 0.86)
- High Speed and High Dynamic Range Video with an Event Camera (TPAMI 2019 · similarity 0.85)
- Event-3DGS: Event-based 3D Reconstruction Using 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.84)
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.
