Research map / Detection and segmentation / Tracking
ModTrack: Sensor-Agnostic Multi-View Tracking via Identity-Informed PHD Filtering with Covariance Propagation
Aditya Iyer, Jack Roberts, Nora Ayanian
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 appearance features to spatial association (joint mode) (ablation) | ModTrack · WildTrack | IDF1: 86.6 -> 95.5 | Improved |
| add appearance features to spatial association (joint mode) (ablation) | ModTrack · MultiviewX | IDF1: 84.2 -> 86.2 | Improved |
| add spatial cues to appearance-only association (ablation) | ModTrack · WildTrack | IDF1: 49.6 -> 95.5 | Improved |
| use chi2 graph clustering instead of RANSAC clustering (ablation) | ModTrack · WildTrack | IDF1: 92.8 -> 95.5 | Improved |
| use precision-weighted fusion instead of simple mean (ablation) | ModTrack · WildTrack | IDF1: 93.0 -> 95.5 | Improved |
| use structured depth-propagated covariance instead of fixed R=0.16I (ablation) | ModTrack · WildTrack | IDF1: 95.0 -> 95.5 | Improved |
| use ground-truth boxes instead of YOLO detections | ModTrack · WildTrack | IDF1: 95.5 -> 97.1 | Improved |
| use ground-truth boxes instead of YOLO detections | ModTrack · MultiviewX | MOTA: 86.3 -> 93.6 | Improved |
| drop cameras from 7 to 5 (ablation) | ModTrack · WildTrack | IDF1: 95.5 -> 85.1 | Hurt |
| drop cameras from 7 to 3 (ablation) | ModTrack · WildTrack | IDF1: 95.5 -> 57.8 | Hurt |
| enlarge Euclidean gate tau_euc from 0.10 to 0.75 (ablation) | ModTrack · MultiviewX | FN clusters / merge errors: 682 to 56; 225 to 481 | Mixed |
| swap radar oracle perception for Doppler perception | ModTrack · RadarScenes | LSTQ: 76.9 -> 42.1 | Hurt |
Closest papers by meaning
- FusionTrack: Collaborative Multi-Object Tracking with Arbitrary Multi-UAVs (ECCV 2026 · similarity 0.89)
- TrackFlow: Multi-Object tracking with Normalizing Flows (ICCV 2023 · similarity 0.89)
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · similarity 0.87)
- Collaborating Vision, Depth, and Thermal Signals for Multi-Modal Tracking: Dataset and Algorithm (NeurIPS 2025 · similarity 0.87)
- More Than Meets the Eye: Enhancing Multi-Object Tracking Even with Prolonged Occlusions (ICML 2025 · similarity 0.87)
- Single-Model and Any-Modality for Video Object Tracking (CVPR 2024 · 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.
