Research map / Video / Video segmentation and tracking
Copy-and-Paste Networks for Deep Video Inpainting
Sungho Lee, Seoung Wug Oh, DaeYeun Won, Seon Joo Kim
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 |
|---|---|---|---|
| apply the context matching network to restore over/under-exposed frames before lane detection | internal color-histogram lane detector · Lane detection, 469 tunnel frames | lane detection accuracy (%): 46.69% -> 83.00% | Improved |
| use Copy-and-Paste Networks instead of optimization-based Huang et al. | video inpainting · DAVIS 2017 (25 videos) | PSNR / SSIM: PSNR 28.14 -> 28.37, SSIM 0.859 -> 0.851 | Mixed |
| use Copy-and-Paste Networks instead of Huang et al. (user study) | video object removal · DAVIS user study | average ranking: 1.74 -> 1.77 | Hurt |
| use Copy-and-Paste Networks instead of VINet | video object removal · DAVIS user study | average ranking: 2.08 -> 1.77 | Improved |
| use feed-forward network instead of optimization (runtime) | Huang et al. · DAVIS object removal | seconds per video: 952 seconds -> 27.14 seconds | Improved |
| add masked softmax in context matching (ablation) | Copy-and-Paste Networks · qualitative DAVIS | sharpness (visual) | Improved |
| add iterative reference frame update with inpainted results (ablation) | Copy-and-Paste Networks · qualitative temporal profile | temporal consistency (visual) | Improved |
Closest papers by meaning
- Deep Learning-Based Image and Video Inpainting: A Survey (IJCV 2024 · similarity 0.90)
- Deep Video Inpainting (CVPR 2019 · similarity 0.89)
- Video Inpainting by Jointly Learning Temporal Structure and Spatial Details (AAAI 2019 · similarity 0.89)
- Recurrent Temporal Aggregation Framework for Deep Video Inpainting (TPAMI 2019 · similarity 0.88)
- Semi-Supervised Video Inpainting With Cycle Consistency Constraints (CVPR 2023 · similarity 0.88)
- Deep Stereo Video Inpainting (CVPR 2023 · similarity 0.88)
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.
