Research map / Video / Video segmentation and tracking
TempFormer: Temporally Consistent Transformer for Video Denoising
Mingyang Song, Yang Zhang, Tunç O. Aydın
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 optical flow warping of RGB inputs (TempFormer++) | TempFormer · DAVIS 480P sigma=10 | PSNR: 39.97 -> 40.17 | Improved |
| add optical flow warping of RGB inputs (TempFormer++) | TempFormer · Set8 sigma=50 | PSNR: 30.96 -> 31.16 | Improved |
| add optical flow warping at higher noise | TempFormer++ · qualitative sigma=30,50 | visual quality | Hurt |
| replace VRT with TempFormer | VRT · DAVIS 480P inference | time per frame (s): 3.54 -> 1.38 | Improved |
| replace VRT with TempFormer | VRT · DAVIS 480P sigma=30 | PSNR: 36.52 -> 35.40 | Hurt |
| add Wavelet decomposition (small model) (ablation) | TempFormer · DAVIS 1080p breakdance sigma=30 | inference time (s/f): 4.21 -> 1.43 | Improved |
| add Wavelet decomposition (small model) (ablation) | TempFormer · DAVIS 1080p breakdance sigma=30 | PSNR: 36.51 -> 36.34 | Hurt |
| use large model (channel 120) with Wavelet instead of small (ablation) | TempFormer · DAVIS 1080p breakdance sigma=30 | PSNR: 36.34 -> 36.59 | Improved |
| use Wavelet Transform instead of Pixel Shuffle (ablation) | TempFormer · DAVIS sigma=30 | PSNR: 35.06 -> 35.10 | Improved |
| add recurrent fine-tuning and Overlap Loss (TCE phase) (ablation) | TempFormer · toy sequence temporal consistency | temporal MAE | Improved |
Closest papers by meaning
- Temporal As a Plugin: Unsupervised Video Denoising with Pre-Trained Image Denoisers (ECCV 2024 · similarity 0.89)
- Instant Video Models: Universal Adapters for Stabilizing Image-Based Networks (NeurIPS 2025 · similarity 0.85)
- Classic Video Denoising in a Machine Learning World: Robust, Fast, and Controllable (CVPR 2025 · similarity 0.85)
- TGBFormer: Transformer-GraphFormer Blender Network for Video Object Detection (AAAI 2025 · similarity 0.85)
- MotionAura: Generating High-Quality and Motion Consistent Videos using Discrete Diffusion (ICLR 2025 · similarity 0.85)
- Improving Interpretation Faithfulness for Vision Transformers (ICML 2024 · similarity 0.85)
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.
