Research map / Detection and segmentation / Tracking
Beyond MOT: Semantic Multi-Object Tracking
Yunhao Li, Qin Li, Hao Wang, Xue Ma, Jiali Yao, Shaohua Dong, Heng Fan, Libo Zhang
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 |
|---|---|---|---|
| train SMOTer end-to-end with semantic tasks vs ByteTrack tracking-only | ByteTrack · BenSMOT test | HOTA: 68.84 -> 71.98 | Improved |
| train SMOTer end-to-end vs ByteTrack | ByteTrack · BenSMOT test | MOTA: 73.87 -> 77.71 | Improved |
| end-to-end SMOTer vs two-stage ByteTrack | ByteTrack · BenSMOT test (video captioning) | CIDEr: 0.266 -> 0.343 | Improved |
| end-to-end SMOTer vs two-stage ByteTrack | ByteTrack · BenSMOT test (interaction) | F1: 0.326 -> 0.368 | Improved |
| use attention-based instead of addition-based video fusion (ablation) | SMOTer VFM · BenSMOT test (video captioning) | CIDEr: 0.127 -> 0.343 | Improved |
| use attention-based instead of MLP-based video fusion (ablation) | SMOTer VFM · BenSMOT test (video captioning) | CIDEr: 0.220 -> 0.343 | Improved |
| use concatenation-based instead of addition-based trajectory fusion (ablation) | SMOTer TFM · BenSMOT test (interaction) | F1: 0.205 -> 0.368 | Improved |
| use concatenation instead of attention for trajectory fusion (ablation) | SMOTer TFM · BenSMOT test (interaction) | F1: 0.364 -> 0.368 | No clear effect |
| raise proposal threshold from 0.3 to 0.5 (ablation) | SMOTer · BenSMOT test | HOTA: 71.98 -> 68.30 | Hurt |
| lower proposal threshold from 0.3 to 0.1 (ablation) | SMOTer · BenSMOT test | HOTA: 71.98 -> 70.51 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- ByteTrack — ByteTrack: Multi-Object Tracking by Associating Every Detection Box (ECCV 2022) · better on 8 of 8 shared results
- MOTRv2 — MOTRv2: Bootstrapping End-to-End Multi-Object Tracking by Pretrained Object Detectors (CVPR 2023) · better on 1 of 1 shared result
- MOTR — MOTR: End-to-End Multiple-Object Tracking with TRansformer (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
- A Multi-modal Global Instance Tracking Benchmark (MGIT): Better Locating Target in Complex Spatio-temporal and Causal Relationship (NeurIPS 2023 · similarity 0.91)
- Referring Multi-Object Tracking (CVPR 2023 · similarity 0.90)
- Referring Multi-Object Tracking with Diverse Dynamic Expressions (TPAMI 2026 · similarity 0.90)
- YesTrack: Referring Multi-Object Tracking via MLLM-based Yes/No Reasoning (ECCV 2026 · similarity 0.90)
- OVT-B: A New Large-Scale Benchmark for Open-Vocabulary Multi-Object Tracking (NeurIPS 2024 · similarity 0.89)
- Multiple Object Tracking as ID Prediction (CVPR 2025 · similarity 0.88)
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.
