Research map / Detection and segmentation / Tracking
Occlusion-Aware SORT: Observing Occlusion for Robust Multi-Object Tracking
Chunjiang Li, Jianbo Ma, Li Shen, Yanru Chen, Liangyin Chen
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 occlusion-aware framework (OAO, BAM, GM) to Hybrid-SORT | Hybrid-SORT · DanceTrack test | HOTA: 62.2 -> 63.1 | Improved |
| plug occlusion-awareness into ByteTrack, OC-SORT, SparseTrack, PD-SORT | four trackers · DanceTrack test | HOTA: 47.1 -> 49.0; 54.6 -> 56.5; 55.5 -> 57.8; 58.2 -> 60.4 | Improved |
| add occlusion-awareness to Hybrid-SORT on SportsMOT | Hybrid-SORT · SportsMOT | HOTA: 73.0 -> 73.4 | Improved |
| add occlusion-awareness to Hybrid-SORT on MOT17 | Hybrid-SORT · MOT17 test | HOTA: 63.6 -> 64.2 HOTA; MOTA 80.6 -> 79.6 | Mixed |
| add OAO to the baseline tracker (ablation) | Hybrid-SORT · DanceTrack val | HOTA: 59.4 -> 59.9 | Improved |
| add BAM to the baseline tracker (ablation) | Hybrid-SORT · DanceTrack val | HOTA: 59.4 -> 60.5 | Improved |
| add Gaussian Map refinement of occlusion coefficient (ablation) | OAO + BAM · DanceTrack val | HOTA: +2.1 HOTA; runtime 24.56 ms per frame | Improved |
| add occlusion-aware modules to Hybrid-SORT with ReID (ablation) | Hybrid-SORT-REID · DanceTrack val | HOTA: 63.1 -> 63.4 | Improved |
| increase tau above about 0.4 in OAO (ablation) | OAO · DanceTrack val | HOTA | Hurt |
| use IoU-based adaptive BAM instead of a constant value (ablation) | BAM · DanceTrack val | HOTA: 61.26 -> 61.49 (best constant 0.4 vs IoU) | Improved |
| add OAO, BAM and Byte to OC-SORT, TrackTrack, SORT (ablation) | OC-SORT, TrackTrack, SORT · DanceTrack val | HOTA: 52.3 -> 53.8; 59.3 -> 59.9; 48.4 -> 50.4 | Improved |
| apply framework when lower body occluded or objects airborne | OA-SORT · DanceTrack sequence 0026 | association performance | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Hybrid-SORT — Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking (AAAI 2024) · better on 7 of 8 shared results
- ByteTrack — ByteTrack: Multi-Object Tracking by Associating Every Detection Box (ECCV 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking (AAAI 2024 · similarity 0.90)
- More Than Meets the Eye: Enhancing Multi-Object Tracking Even with Prolonged Occlusions (ICML 2025 · similarity 0.89)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.87)
- OVT-B: A New Large-Scale Benchmark for Open-Vocabulary Multi-Object Tracking (NeurIPS 2024 · similarity 0.87)
- SMILEtrack: SiMIlarity LEarning for Occlusion-Aware Multiple Object Tracking (AAAI 2024 · similarity 0.86)
- Sampling-Resilient Multi-Object Tracking (AAAI 2024 · 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.
