Research map / 3D vision / Human pose estimation
Multiple People Tracking by Lifted Multicut and Person Re-Identification
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 lifted edges instead of only regular edges | MP tracking model · MOT16 test set | MOTA: +2.4% | Helped |
| add body-part score maps to the image-pair network (ablation) | StackNet person re-identification · MOT16-02 and MOT16-11 re-identification test pairs | verification accuracy: 86.9% -> 90.0% | Helped |
| stack the image pair along the RGB channel (ablation) | SiameseNet person re-identification · MOT16-02 and MOT16-11 re-identification test pairs | verification accuracy: 84.7% -> 86.9% | Helped |
| use a Siamese architecture instead of ID-Net (ablation) | person re-identification network · MOT16-02 and MOT16-11 re-identification test pairs | verification accuracy: 80.4% -> 84.7% | Helped |
| use lifted edges rather than only regular edges (ablation) | MP tracking model · MOT16-11 validation, δmax=150 | MOTA: 32.1 -> 51.1 (+19.1) | Helped |
| use lifted edges rather than only regular edges (ablation) | MP tracking model · MOT16-02 validation, δmax=120 | MOTA: 19.1 -> 22.3 (+3.2) | Helped |
| extend regular edges to long temporal distances without lifted edges (ablation) | MP tracking model · MOT16-11 validation, δmax=150 | MOTA: 54.2 -> 32.1 | Hurt |
| combine spatio-temporal, dense matching, and Re-ID features instead of using an individual feature (ablation) | pairwise affinity model · MOT16-02 and MOT16-11 | pairwise verification accuracy | Helped |
Closest papers by meaning
- PoseTrack21: A Dataset for Person Search, Multi-Object Tracking and Multi-Person Pose Tracking (CVPR 2022 · similarity 0.89)
- PoseTrack: Joint Multi-Person Pose Estimation and Tracking (CVPR 2017 · similarity 0.88)
- ArtTrack: Articulated Multi-Person Tracking in the Wild (CVPR 2017 · similarity 0.88)
- DeepCut: Joint Subset Partition and Labeling for Multi Person Pose Estimation (CVPR 2016 · similarity 0.88)
- Combining Detection and Tracking for Human Pose Estimation in Videos (CVPR 2020 · similarity 0.87)
- CoMotion: Concurrent Multi-person 3D Motion (ICLR 2025 · 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.