Research map / Video / Video segmentation and tracking
Following Motion for Sequential Modeling in Video Frame Interpolation
JaeHyun Park, Nam Ik Cho
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 modified EDFFN block to SSM baseline (ablation) | SSM baseline · Vimeo-90K | PSNR (dB): 35.96 -> 36.10 | Improved |
| add Motion-Guided Serialization (ablation) | SSM + EDFFN · SNU-FILM Hard | PSNR (dB): 29.80 -> 30.24 | Improved |
| add Contextual Synthesis (ablation) | SSM + EDFFN + MGS · SNU-FILM Hard | PSNR (dB): 30.24 -> 31.04 | Improved |
| add Contextual Synthesis (ablation) | SSM + EDFFN + MGS · UCF101 | PSNR (dB): 35.56 -> 35.53 | No clear effect |
| add Motion-Guided Serialization (ablation) | SSM + EDFFN · Vimeo-90K | PSNR (dB): 36.10 -> 36.26 | Improved |
| use motion-guided scan instead of interleaved scan (ablation) | VFI Mamba · Vimeo90K (100 images), serialized sequence | Sequential Feature Similarity: 0.46 (31% above next best, Interleaved) | Improved |
| use RAFT-Small instead of RAFT-Large flow (ablation) | MGMVFI · SNU-FILM Hard | PSNR (dB): 31.04 -> 30.92 | No clear effect |
| use PWC-Net instead of RAFT-Large flow (ablation) | MGMVFI · SNU-FILM Hard | PSNR (dB): 31.04 -> 30.70 | Hurt |
| MGMVFI vs VFIMamba on motion extremes | MGMVFI · SNU-FILM Extreme | PSNR (dB): 25.79 -> 25.86 | Improved |
| MGMVFI vs VFIMamba on easy/medium splits | MGMVFI · SNU-FILM Easy | PSNR (dB): 40.51 -> 40.45 | 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.
- VFIMamba — VFIMamba: Video Frame Interpolation with State Space Models (NeurIPS 2024) · better on 5 of 8 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
- VFIMamba: Video Frame Interpolation with State Space Models (NeurIPS 2024 · similarity 0.91)
- LC-Mamba: Local and Continuous Mamba with Shifted Windows for Frame Interpolation (CVPR 2025 · similarity 0.89)
- Disentangled Motion Modeling for Video Frame Interpolation (AAAI 2025 · similarity 0.88)
- Generalizable Implicit Motion Modeling for Video Frame Interpolation (NeurIPS 2024 · similarity 0.88)
- BiM-VFI: Bidirectional Motion Field-Guided Frame Interpolation for Video with Non-uniform Motions (CVPR 2025 · similarity 0.88)
- Sparse Global Matching for Video Frame Interpolation with Large Motion (CVPR 2024 · similarity 0.87)
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.
