Research map / Video / Video segmentation and tracking
AMT: All-Pairs Multi-Field Transforms for Efficient Frame Interpolation
Zhen Li, Zuo-Liang Zhu, Ling-Hao Han, Qibin Hou, Chun-Le Guo, Ming-Ming Cheng
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 |
|---|---|---|---|
| Use AMT-S rather than IFRNet-B | Vimeo90K · Vimeo90K | PSNR: 35.80 -> 35.97 | Improved |
| Use AMT-S rather than IFRNet-B | SNU-FILM Hard · SNU-FILM Hard | PSNR: 30.41 -> 30.60 | Improved |
| Use AMT-L rather than IFRNet-L | Vimeo90K · Vimeo90K | PSNR: 36.20 -> 36.35 | Improved |
| Use AMT-G rather than VFIFormer | Vimeo90K · Vimeo90K | PSNR: 36.50 -> 36.53 | Improved |
| Add a dedicated correlation encoder (ablation) | AMT-S · Vimeo90K | PSNR: 35.76 -> 35.97 | Improved |
| Use bidirectional rather than unidirectional correlation volumes (ablation) | AMT-S · Vimeo90K | PSNR: 35.93 -> 35.97 | Improved |
| Use all-pairs rather than PWC-like correlation volumes (ablation) | AMT-S · Vimeo90K | PSNR: 35.61 -> 35.97 | Improved |
| Scale flows for correlation lookup instead of RAFT-style lookup (ablation) | AMT-S · Vimeo90K | PSNR: 35.93 -> 35.97 | Improved |
| Initialize scaled lookup with bilateral flows rather than zero (ablation) | AMT-S · SNU-FILM Hard | PSNR: 30.56 -> 30.60 | Improved |
| Use intermediate content features as guidance rather than visible-frame features (ablation) | AMT-S · Vimeo90K | PSNR: 35.95 -> 35.97 | Improved |
| Update content features jointly with flows (ablation) | AMT-S · SNU-FILM Extreme | PSNR: 25.22 -> 25.30 | Improved |
| Use cross-scale rather than single-scale updates (ablation) | AMT-S · SNU-FILM Extreme | PSNR: 25.22 -> 25.30 | Improved |
Closest papers by meaning
- Video Frame Interpolation With Many-to-Many Splatting and Spatial Selective Refinement (TPAMI 2023 · similarity 0.88)
- Many-to-Many Splatting for Efficient Video Frame Interpolation (CVPR 2022 · similarity 0.88)
- AdaCoF: Adaptive Collaboration of Flows for Video Frame Interpolation (CVPR 2020 · similarity 0.87)
- Context-Aware Synthesis for Video Frame Interpolation (CVPR 2018 · similarity 0.87)
- ST-MFNet: A Spatio-Temporal Multi-Flow Network for Frame Interpolation (CVPR 2022 · similarity 0.87)
- Super SloMo: High Quality Estimation of Multiple Intermediate Frames for Video Interpolation (CVPR 2018 · 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.
