Research map / Video / Video segmentation and tracking
Efficient Multi-Stage Video Denoising With Recurrent Spatio-Temporal Fusion
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 previous refined output instead of previous fused image for recurrence (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 42.38 | Hurt |
| Remove current noisy low-pass input from the denoising stage (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 41.87 | Hurt |
| Remove the color transform (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 42.36 | Hurt |
| Replace the learnable frequency transform with pixel shuffling (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 42.35 | Hurt |
| Remove refinement and increase denoising capacity to keep overall complexity (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 42.46 | Hurt |
| Remove variance input from all networks (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 41.39 | Hurt |
| Disable fusion and refinement (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 41.35 | Hurt |
| Increase capacity to FCNN 4/64, DCNN 4/512, RCNN 4/64 (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 44.51 | Helped |
| Use FCNN 4/64, DCNN 6/256, RCNN 4/64 (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 44.48 | Helped |
| Use FCNN 3/64, DCNN 3/64, RCNN 3/64 (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 43.57 | Helped |
| Use FCNN 4/16, DCNN 5/64, RCNN 3/64 (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 43.83 | Helped |
| Use FCNN 4/16, DCNN 6/64, RCNN 1/32 (ablation) | EMVD · CRVD raw | PSNR: 42.63 -> 44.05 | Helped |
Closest papers by meaning
- FastDVDnet: Towards Real-Time Deep Video Denoising Without Flow Estimation (CVPR 2020 · similarity 0.88)
- ReMoNet: Recurrent Multi-Output Network for Efficient Video Denoising (AAAI 2022 · similarity 0.87)
- Classic Video Denoising in a Machine Learning World: Robust, Fast, and Controllable (CVPR 2025 · similarity 0.86)
- Denoising as Adaptation: Noise-Space Domain Adaptation for Image Restoration (ICLR 2025 · similarity 0.83)
- Masked and Shuffled Blind Spot Denoising for Real-World Images (CVPR 2024 · similarity 0.83)
- APR-RD: Complemental Two Steps for Self-Supervised Real Image Denoising (AAAI 2025 · similarity 0.83)
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.