Research map / Video / Video segmentation and tracking
Let's See Clearly: Contaminant Artifact Removal for Moving Cameras
Xiaoyu Li, Bo Zhang, Jing Liao, Pedro V. Sander
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 attention map detection (ablation) | single-frame restoration · Synthetic dirty-lens test set (30 clips) | PSNR: 33.70 -> 35.37 | Improved |
| add flow completion (ablation) | single-frame restoration · Synthetic dirty-lens test set (30 clips) | PSNR: 34.09 -> 35.37 | Improved |
| add spatial restoration module (ablation) | single-frame restoration · Synthetic dirty-lens test set (30 clips) | PSNR: 29.61 -> 35.37 | Improved |
| add spatial restoration module (ablation) | single-frame restoration · Synthetic dirty-lens test set (30 clips) | SSIM: 0.953 -> 0.980 | Improved |
| add flow completion (ablation) | single-frame restoration · Synthetic dirty-lens test set (30 clips) | Warping error E_warp: 0.0046 -> 0.0045 | No clear effect |
| replace plain conv with feature fusion layer in flow completion (ablation) | flow completion network · Flow completion, synthetic test set | EPE: 2.10 -> 1.72 | Improved |
| replace gated conv with feature fusion layer (ablation) | flow completion network · Flow completion, synthetic test set | EPE: 1.89 -> 1.72 | Improved |
| replace decoder with learned upsampling module (ablation) | flow completion network · Flow completion, synthetic test set | EPE: 1.72 -> 1.60 | Improved |
| add flow completion network (vs degraded input flow) (ablation) | flow estimation · Flow completion, synthetic test set | EPE: 3.17 -> 1.60 | Improved |
| add multi-frame processing stage (ablation) | two-stage method · Synthetic dirty-lens test set (30 clips) | Warping error E_warp: 0.0045 -> 0.0035 | Improved |
| add multi-frame processing stage (ablation) | two-stage method · Synthetic dirty-lens test set (30 clips) | PSNR: 35.37 -> 34.98 | Hurt |
| add multi-frame processing stage (ablation) | two-stage method · Synthetic dirty-lens test set (30 clips) | SSIM: 0.980 -> 0.979 | Hurt |
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.
- FastDVDnet — FastDVDnet: Towards Real-Time Deep Video Denoising Without Flow Estimation (CVPR 2020) · better on 3 of 3 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
- Precise Object and Effect Removal with Adaptive Target-Aware Attention (CVPR 2026 · similarity 0.84)
- DeLux: Cross-Modal Local Artifact Restoration in Video Using Neuromorphic Data (ECCV 2026 · similarity 0.84)
- EffectErase: Joint Video Object Removal and Insertion for High-Quality Effect Erasing (CVPR 2026 · similarity 0.84)
- Decoupling Degradations with Recurrent Network for Video Restoration in Under-Display Camera (AAAI 2024 · similarity 0.83)
- Visual-Instructed Degradation Diffusion for All-in-One Image Restoration (CVPR 2025 · similarity 0.83)
- Making Old Film Great Again: Degradation-aware State Space Model for Old Film Restoration (CVPR 2025 · similarity 0.82)
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.
