Research map / 3D vision / Depth and stereo
Spike Stream Memory Transfer for Dynamic Scene Reconstruction
Yanchen Dong, Ruiqin Xiong, Rui Zhao, Xinfeng Zhang, Tiejun Huang
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 SSMT on top of WGSE | WGSE · REDS-SPK | PSNR (dB): 39.43 -> 40.28 | Improved |
| add SSMT on top of Spk2ImgNet | Spk2ImgNet · REDS-SPK | PSNR (dB): 40.13 -> 40.37 | Improved |
| add SSMT on top of STIR | STIR · DAVIS-SPK | PSNR (dB): 35.33 -> 36.45 | Improved |
| add SSMT on top of Spk2ImgNet | Spk2ImgNet · Real-SPK | NIQE: 5.5243 -> 4.4535 | Improved |
| add SSMT on top of WGSE | WGSE · Running time | seconds: 0.8819 -> 0.3906 | Improved |
| split input spike clip into subclips (ablation) | reconstruction · REDS-SPK | PSNR (dB): 39.60 -> 39.90 | Improved |
| add DCN-based alignment of split features (ablation) | reconstruction · REDS-SPK | PSNR (dB): 39.90 -> 40.00 | Improved |
| replace DCN alignment with spike interval-guided alignment (SIA) (ablation) | reconstruction · REDS-SPK | PSNR (dB): 40.00 -> 40.09 | Improved |
| add light intensity memory transfer (ablation) | reconstruction · REDS-SPK | PSNR (dB): 40.09 -> 40.34 | Improved |
| add motion memory transfer (ablation) | motion estimation · REDS-SPK | PSNR (dB): 40.16 -> 40.26 | Improved |
| add learnable flow head (ablation) | motion estimation · REDS-SPK | PSNR (dB): 40.26 -> 40.34 | Improved |
| use spike intervals instead of spike frames for motion estimation (ablation) | motion estimation · REDS-SPK | PSNR (dB): 40.27 -> 40.34 | 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.
- Spk2ImgNet — Spk2ImgNet: Learning To Reconstruct Dynamic Scene From Continuous Spike Stream (CVPR 2021) · better on 4 of 4 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
- SpikeReveal: Unlocking Temporal Sequences from Real Blurry Inputs with Spike Streams (NeurIPS 2024 · similarity 0.90)
- Rethinking High-speed Image Reconstruction Framework with Spike Camera (AAAI 2025 · similarity 0.89)
- Finding Visual Saliency in Continuous Spike Stream (AAAI 2024 · similarity 0.89)
- Spike-guided Motion Deblurring with Unknown Modal Spatiotemporal Alignment (CVPR 2024 · similarity 0.89)
- Boosting Spike Camera Image Reconstruction from a Perspective of Dealing with Spike Fluctuations (CVPR 2024 · similarity 0.88)
- Spk2SRImgNet: Super-Resolve Dynamic Scene from Spike Stream via Motion Aligned Collaborative Filtering (CVPR 2025 · 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.
