Research map / Video / Video segmentation and tracking
Learning Image Matching by Simply Watching Video
Gucan Long, Laurent Kneip, Jose M. Alvarez, Hongdong Li, Xiaohu Zhang, Qifeng Yu
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 |
|---|---|---|---|
| initialize DeepFlow with MIND matches instead of no matches | DeepFlow · KITTI 2012 flow training | APE: 3.55 -> 2.89 | Improved |
| initialize DeepFlow with MIND matches instead of no matches | DeepFlow · MPI-Sintel flow training | APE: 6.63 -> 5.78 | Improved |
| initialize DeepFlow with DeepMatching instead of MIND matches | DeepFlow · KITTI 2012 flow training | APE: 2.89 -> 2.63 | Improved |
| initialize DeepFlow with MIND matches instead of HoG | DeepFlow · MPI-Sintel flow training | APE: 5.46 -> 5.78 | Hurt |
| initialize DeepFlow with MIND matches instead of HoG | DeepFlow · KITTI 2012 flow training | APE: 3.06 -> 2.89 | Improved |
| use MIND matches instead of KLT matching | matching · KITTI 2012 flow training (matching) | Accuracy@10: 0.826 -> 0.915 | Improved |
| use MIND matches instead of HoG matching | matching · KITTI 2012 flow training (matching) | Accuracy@10: 0.805 -> 0.915 | Improved |
| use MIND matches instead of KLT matching | matching · MPI-Sintel training Final (matching) | Accuracy@5: 0.808 -> 0.719 | Hurt |
| use MIND matches instead of HoG matching | matching · MPI-Sintel training Final (matching) | Accuracy@10: 0.875 -> 0.876 | No clear effect |
| use DeepMatching instead of MIND | matching · MPI-Sintel training Final (matching) | APE: 5.838 -> 3.240 | Improved |
| use Charbonnier loss instead of L2 loss (ablation) | MIND frame interpolation network · KITTI/Sintel training | interpolation quality | Improved |
| use MIND instead of DeepFlow for interpolation | frame interpolation · KITTI | mean interpolation error: 33.30 -> 30.99 | Improved |
Closest papers by meaning
- Rethinking Self-Supervised Correspondence Learning: A Video Frame-Level Similarity Perspective (ICCV 2021 · similarity 0.87)
- Convolutional Hough Matching Networks (CVPR 2021 · similarity 0.87)
- Image Matching from Handcrafted to Deep Features: A Survey (IJCV 2020 · similarity 0.87)
- Learning Cross-Video Neural Representations for High-Quality Frame Interpolation (ECCV 2022 · similarity 0.87)
- Convolutional Neural Network Architecture for Geometric Matching (CVPR 2017 · similarity 0.86)
- A Temporally-Aware Interpolation Network for Video Frame Inpainting (TPAMI 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.
