Research map / Video / Video segmentation and tracking
Temporally Robust Global Motion Compensation by Keypoint-Based Congealing
S. Morteza Safdarnejad, Yousef Atoum, Xiaoming Liu
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 joint keypoint-based congealing (TRGMC) instead of sequential GMC (RGMC) | Global motion compensation · Quantitative GMC dataset | average BRE: 0.097 -> 0.058 | Improved |
| use TRGMC instead of MLESAC | Global motion compensation · Quantitative GMC dataset | average BRE: 0.116 -> 0.058 | Improved |
| use backward-forward links instead of backward-only links (ablation) | TRGMC · Quantitative GMC dataset | average BRE: 0.060 -> 0.058 | Improved |
| use backward-forward links instead of backward-only links (ablation) | TRGMC · Quantitative GMC dataset | processing time (s/frame): 0.41 -> 0.64 | Hurt |
| use TRGMC instead of HEASK | Global motion compensation · Quantitative GMC dataset | processing time (s/frame): 7.47 -> 0.64 | Improved |
| use TRGMC instead of RGMC | Global motion compensation · Quantitative GMC dataset | processing time (s/frame): 3.47 -> 0.64 | Improved |
| lower keyframe step to process more frames in joint stage (ablation) | TRGMC · 5-video subset | BRE and efficiency | Mixed |
| apply TRGMC camera motion removal before multi-object tracking | Multi-object tracking · Football video, 566 frames | MOTA (%): 63.79% -> 84.23% | Improved |
Closest papers by meaning
- UCMCTrack: Multi-Object Tracking with Uniform Camera Motion Compensation (AAAI 2024 · similarity 0.84)
- Sparse Global Matching for Video Frame Interpolation with Large Motion (CVPR 2024 · similarity 0.83)
- MoDec-GS: Global-to-Local Motion Decomposition and Temporal Interval Adjustment for Compact Dynamic 3D Gaussian Splatting (CVPR 2025 · similarity 0.82)
- MotionPro: A Precise Motion Controller for Image-to-Video Generation (CVPR 2025 · similarity 0.82)
- HumanMM: Global Human Motion Recovery from Multi-shot Videos (CVPR 2025 · similarity 0.82)
- TrajLoc: Trajectory-Attention Localization for Multi-Object Motion Control (NeurIPS 2026 · similarity 0.82)
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.
