Research map / Video / Video segmentation and tracking
Unsupervised Deep Video Denoising
Dev Yashpal Sheth, Sreyas Mohan, Joshua L. Vincent, Ramon Manzorro, Peter A. Crozier, Mitesh M. Khapra, Eero P. Simoncelli, Carlos Fernandez-Granda
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 5 input frames instead of 1 (ablation) | UDVD · DAVIS test (sigma=30) | PSNR: 32.80 -> 33.92 | Improved |
| use 5 input frames instead of 1 (ablation) | UDVD · Set8 (sigma=30) | PSNR: 30.91 -> 32.01 | Improved |
| use 5 input frames instead of 3 (ablation) | UDVD · DAVIS test (sigma=30) | PSNR: 33.48 -> 33.92 | Improved |
| use 5 input frames instead of 1 at high noise (ablation) | UDVD · DAVIS test (sigma=80) | PSNR: 28.37 -> 29.63 | Improved |
| add early stopping on held-out last 5 frames (ablation) | UDVD-S · 9 videos (sigma=90) | PSNR: 23.33 -> 27.60 | Improved |
| add early stopping on held-out last 5 frames (ablation) | UDVD-S · 9 videos (sigma=30) | PSNR: 30.80 -> 32.39 | Improved |
| add spatial flipping and time reversal augmentation (with early stopping) (ablation) | UDVD-S · 9 videos (sigma=30) | PSNR: 32.39 -> 32.61 | Improved |
| add spatial flipping and time reversal augmentation (with early stopping) (ablation) | UDVD-S · 9 videos (sigma=90) | PSNR: 27.60 -> 27.80 | Improved |
| add flip and time reversal augmentation (without early stopping) (ablation) | UDVD-S · 9 videos (sigma=30) | PSNR: 30.80 -> 32.24 | Improved |
| train UDVD on single video (UDVD-S) instead of DAVIS dataset | UDVD · Set8 individual-video (sigma=30) | PSNR: 31.90 -> 32.90 | Improved |
| train UDVD-S without pretraining instead of fine-tuning MF2F | UDVD-S · Set8 individual-video (sigma=30) | PSNR: 31.84 -> 32.90 | Improved |
| unsupervised UDVD instead of supervised RViDeNet on raw video | UDVD · Raw video dataset | PSNR: 43.97 -> 44.69 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- FastDVDnet — FastDVDnet: Towards Real-Time Deep Video Denoising Without Flow Estimation (CVPR 2020) · better on 2 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
- FastDVDnet: Towards Real-Time Deep Video Denoising Without Flow Estimation (CVPR 2020 · similarity 0.90)
- Unsupervised Deep Video Denoising with Untrained Network (AAAI 2023 · similarity 0.90)
- Unsupervised Image Restoration Using Partially Linear Denoisers (TPAMI 2021 · similarity 0.89)
- Universal Denoising Networks : A Novel CNN Architecture for Image Denoising (CVPR 2018 · similarity 0.89)
- Classic Video Denoising in a Machine Learning World: Robust, Fast, and Controllable (CVPR 2025 · similarity 0.88)
- Efficient Online Training for Zero-Shot Time-Lapse Microscopy Denoising and Super-Resolution (AAAI 2025 · similarity 0.86)
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.
