Research map / Video / Video segmentation and tracking
Deep Slow Motion Video Reconstruction With Hybrid Imaging System
Avinash Paliwal, Nima Khademi Kalantari
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 |
|---|---|---|---|
| add visibility maps (ablation) | appearance estimation network · 30fps main, 240fps 270p aux | LPIPS: 0.1193 -> 0.1110 | Improved |
| add contextual information (ablation) | appearance estimation network · 30fps main, 240fps 270p aux | LPIPS: 0.1110 -> 0.1012 | Improved |
| raise auxiliary resolution from 180p to 540p (ablation) | full system · aux resolution sweep | LPIPS: 0.1065 -> 0.1005 | No clear effect |
| lower auxiliary resolution from 180p to 45p (ablation) | full system · aux resolution sweep | LPIPS: 0.1065 -> 0.2011 | Hurt |
| raise main video frame rate from 24 to 120 fps (ablation) | full system · main frame rate sweep | LPIPS: 0.1095 -> 0.0535 | Improved |
| change gamma of auxiliary video from 1.00 to 0.65 (ablation) | full system · gamma robustness | LPIPS: 0.1005 -> 0.1016 | No clear effect |
| add Gaussian noise sigma 75 to auxiliary video (ablation) | full system · noise robustness | LPIPS: 0.1005 -> 0.2747 | Hurt |
| denoise noisy auxiliary video with VBM4D at sigma 75 (ablation) | full system · noise robustness | LPIPS: 0.2747 -> 0.1380 | Improved |
| desynchronize auxiliary video by 3 frames (ablation) | full system · desynchronization | LPIPS: 0.1005 -> 0.1326 | Hurt |
| use hybrid auxiliary video instead of linear-motion interpolation (Ours vs Super SloMo) | full system · Slow Flow | LPIPS: 0.1525 -> 0.1332 | Improved |
Closest papers by meaning
- FILM: Frame Interpolation for Large Motion (ECCV 2022 · similarity 0.88)
- Learning to Extract Flawless Slow Motion From Blurry Videos (CVPR 2019 · similarity 0.87)
- AdaCoF: Adaptive Collaboration of Flows for Video Frame Interpolation (CVPR 2020 · similarity 0.86)
- Video Frame Interpolation without Temporal Priors (NeurIPS 2020 · similarity 0.86)
- Zooming Slow-Mo: Fast and Accurate One-Stage Space-Time Video Super-Resolution (CVPR 2020 · similarity 0.86)
- Depth-Aware Video Frame Interpolation (CVPR 2019 · 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.
