Research map / Video / Video segmentation and tracking
Video Frame Interpolation via Adaptive Convolution
Check 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 adaptive-convolution interpolation instead of optical-flow-based interpolation | Middlebury testing set: Baske. · Middlebury testing set: Baske. | average interpolation error: 6.13 -> 5.33 | Helped |
| Use adaptive-convolution interpolation instead of optical-flow-based interpolation | Middlebury testing set: Dumpt. · Middlebury testing set: Dumpt. | average interpolation error: 7.36 -> 7.30 | Helped |
| Use adaptive-convolution interpolation instead of optical-flow-based interpolation | Middlebury testing set: Everg. · Middlebury testing set: Everg. | average interpolation error: 7.75 -> 6.94 | Helped |
| Use adaptive-convolution interpolation instead of optical-flow-based interpolation | Middlebury testing set: Backy. · Middlebury testing set: Backy. | average interpolation error: 10.2 -> 10.2 | No clear effect |
| Use adaptive-convolution interpolation instead of the strongest listed baseline | Middlebury testing set: Mequ., Schef., Urban, Teddy · Middlebury testing set | average interpolation error: 2.89 -> 3.57; 3.47 -> 4.34; 3.62 -> 5.00; 5.20 -> 6.91 | Hurt |
| Add gradient loss to color loss (ablation) | Adaptive Convolution · video frame interpolation | sharpness | Helped |
| Replace convolution-kernel prediction with direct pixel synthesis (ablation) | Adaptive Convolution network · Sintel example | visual sharpness | Hurt |
| Use shift-and-stitch instead of pixel-wise inference (ablation) | Adaptive Convolution · 1280 × 720 frame | inference time: 104 seconds -> 9 seconds | Helped |
| Use the current kernel size for motion beyond 41 pixels | Adaptive Convolution · large-motion interpolation | interpolation quality | Hurt |
Closest papers by meaning
- Channel Attention Is All You Need for Video Frame Interpolation (AAAI 2020 · similarity 0.92)
- Depth-Aware Video Frame Interpolation (CVPR 2019 · similarity 0.92)
- Super SloMo: High Quality Estimation of Multiple Intermediate Frames for Video Interpolation (CVPR 2018 · similarity 0.91)
- Disentangled Motion Modeling for Video Frame Interpolation (AAAI 2025 · similarity 0.87)
- Generative Inbetweening: Adapting Image-to-Video Models for Keyframe Interpolation (ICLR 2025 · similarity 0.86)
- EDEN: Enhanced Diffusion for High-quality Large-motion Video Frame Interpolation (CVPR 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.