Research map / Video / Video segmentation and tracking
Exploring Motion Ambiguity and Alignment for High-Quality Video Frame Interpolation
Read the paperCheck 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 motion ambiguity-aware texture consistency loss alongside L1 supervision (ablation) | Baseline VFI model · Vimeo-Triplets-Test | PSNR: 35.90 -> 36.21; +0.31 | Helped |
| Add cross-scale pyramid alignment (ablation) | Baseline VFI model · Vimeo-Triplets-Test | PSNR: 35.90 -> 36.56; +0.66 | Helped |
| Use both TCL and CSPA (ablation) | Full VFI model · Vimeo-Triplets-Test | PSNR: 35.90 -> 36.76; +0.86 | Helped |
| Set the TCL weight alpha to 0.1 rather than 0 (ablation) | TCL VFI model · Vimeo-Triplets-Test | PSNR: 36.56 -> 36.76 | Helped |
| Increase TCL weight alpha from 0.1 to 2.0 (ablation) | TCL VFI model · Vimeo-Triplets-Test | PSNR: 36.76 -> 36.54 | Hurt |
| Set TCL weight alpha to 10.0 (ablation) | TCL VFI model · Vimeo-Triplets-Test | PSNR | Hurt |
| Increase TCL patch size from 3 to 7 (ablation) | TCL VFI model · Vimeo-Triplets-Test | PSNR: 36.76 -> 36.56 | Hurt |
| Use RGB rather than census-space patch matching for TCL (ablation) | TCL VFI model · Vimeo-Triplets-Test | PSNR: 36.76 -> 36.57 | Hurt |
| Use the proposed model instead of SoftSplat | Single-frame VFI · Vimeo-Triplets-Test | PSNR: 36.10 -> 36.76; 0.66dB | Helped |
| Use the proposed model instead of QVI | Multi-frame VFI · Vimeo90K-7f | PSNR: 35.19 -> 36.50 | Helped |
| Use the proposed extrapolation model instead of FLAVR | Video frame extrapolation · Vimeo-Triplets-Test | PSNR: 31.14 -> 32.05; 0.91dB | Helped |
| Add two VFI-generated intermediate frames to the BasicVSR input | BasicVSR · Video super-resolution | PSNR: 26.65 -> 26.75; ↑ 0.1 | Helped |
Closest papers by meaning
- Perception-Oriented Video Frame Interpolation via Asymmetric Blending (CVPR 2024 · similarity 0.89)
- AdaCoF: Adaptive Collaboration of Flows for Video Frame Interpolation (CVPR 2020 · similarity 0.89)
- Exploring Discontinuity for Video Frame Interpolation (CVPR 2023 · similarity 0.88)
- Disentangled Motion Modeling for Video Frame Interpolation (AAAI 2025 · similarity 0.88)
- Sparse Global Matching for Video Frame Interpolation with Large Motion (CVPR 2024 · similarity 0.87)
- VFIMamba: Video Frame Interpolation with State Space Models (NeurIPS 2024 · 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.