Research map / Video / Video segmentation and tracking
Spk2VidNet: A Hierarchical Recurrent Architecture for High-Fidelity Video Reconstruction from Long Spike-Camera Streams
Yuanlin Wang, Ruiqin Xiong, Jiyu Xie, Zhenkun Zhu, Zhaofei Yu, Xiaopeng Fan, 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 MFCA module to baseline (2 layers) (ablation) | Spk2VidNet baseline · REDS-LSSR (x4), full sequence | PSNR: 28.39 -> 29.41 | Improved |
| add CMF module to baseline (2 layers) (ablation) | Spk2VidNet baseline · REDS-LSSR (x4), full sequence | PSNR: 28.39 -> 29.27 | Improved |
| combine MFCA and CMF (2 layers) (ablation) | Spk2VidNet baseline · REDS-LSSR (x4), full sequence | PSNR: 28.39 -> 29.79 | Improved |
| use independent alignment instead of MFCA (2 layers) (ablation) | Spk2VidNet · REDS-LSSR (x4), full sequence | PSNR: 29.41 -> 29.02 | Hurt |
| increase propagation layers from 1 to 2 (ablation) | Spk2VidNet full · REDS-LSSR (x4), full sequence | PSNR: 29.66 -> 29.79 | Improved |
| add second propagation layer to baseline without MFCA/CMF (ablation) | Spk2VidNet baseline · REDS-LSSR (x4), full sequence | PSNR: 28.37 -> 28.39 | No clear effect |
| add segment-wise training with state transfer (ablation) | Spk2VidNet · REDS-LSSR (x4), full sequence | PSNR: 29.58 -> 29.79 | Improved |
| add segment-wise training with state transfer (ablation) | Spk2VidNet · REDS-LSSR (x4), full sequence | SSIM: 0.8377 -> 0.8432 | Improved |
| use Spk2VidNet instead of SCSRNet | SCSR methods · Adobe240-LSSR (x4 SR) | PSNR: 31.31 -> 32.36 | Improved |
| use Spk2VidNet instead of SCSRNet | SCSR methods · REDS-LSSR (x8 SR) | PSNR: 25.81 -> 26.20 | Improved |
| use Spk2VidNet instead of SCSRNet for runtime | SCSR methods · 180x320x461 input (x4) | runtime (ms): 187 -> 43 | Improved |
Closest papers by meaning
- Spk2SRImgNet: Super-Resolve Dynamic Scene from Spike Stream via Motion Aligned Collaborative Filtering (CVPR 2025 · similarity 0.92)
- Spk2ImgNet: Learning To Reconstruct Dynamic Scene From Continuous Spike Stream (CVPR 2021 · similarity 0.91)
- Spatio-Temporal Interactive Learning for Efficient Image Reconstruction of Spiking Cameras (NeurIPS 2024 · similarity 0.91)
- Rethinking High-speed Image Reconstruction Framework with Spike Camera (AAAI 2025 · similarity 0.89)
- SpikeReveal: Unlocking Temporal Sequences from Real Blurry Inputs with Spike Streams (NeurIPS 2024 · similarity 0.88)
- Finding Visual Saliency in Continuous Spike Stream (AAAI 2024 · similarity 0.87)
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.
