Research map / Detection and segmentation / Tracking
SiamMOT: Siamese Multi-Object Tracking
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 |
|---|---|---|---|
| Add an instance-level motion tracker to Faster-RCNN | Faster-RCNN on MOT17 train · MOT17 train | MOTA: 58.6 -> 63.3 | Helped |
| Use EMM instead of IMM (ablation) | SiamMOT · MOT17 train and TAO-person | MOTA / TAP@0.5: 61.5 -> 63.3; 34.7% -> 35.3% | Helped |
| Replace Tracktor motion with flow-based motion (ablation) | Faster-RCNN (Tracktor) · Caltech Roadside Pedestrians | MOTA: 15.9 -> 41.8 | Helped |
| Train using positive, hard, and negative triplets rather than positive and hard triplets only (ablation) | SiamMOT · MOT17 | MOTA: 59.7 -> 63.3 | Helped |
| Train using positive, hard, and negative triplets rather than positive and negative triplets only (ablation) | SiamMOT · MOT17 | MOTA: 62.7 -> 63.3 | Helped |
| Use positive and negative triplets instead of positive and hard triplets (ablation) | SiamMOT · Caltech Roadside Pedestrians | IDF1: 81.3 -> 79.0 | Hurt |
| Jointly train the tracker with the detection network (ablation) | SiamMOT EMM · MOT17 | MOTA: 61.5 -> 63.3 | Helped |
| Jointly train the detector with the tracker (ablation) | Faster-RCNN · MOT17 | AP@IOU=0.5: 73.4% -> 73.3% | No clear effect |
| Keep tracks active during short periods of invisibility (ablation) | SiamMOT · MOT17 | IDF1: 58.4 -> 60.8 | Helped |
| Extend the track-retention period from 15 to 60 frames (ablation) | SiamMOT · MOT17 | IDF1: 60.8 -> 60.2 | Hurt |
| Add person re-identification tracklet linking (ablation) | SiamMOT with DLA-169 · TAO-person validation | TAP@0.5: 42.1% -> 44.3% | Helped |
Closest papers by meaning
- High Performance Visual Tracking With Siamese Region Proposal Network (CVPR 2018 · similarity 0.91)
- SiamRPN++: Evolution of Siamese Visual Tracking With Very Deep Networks (CVPR 2019 · similarity 0.90)
- Siamese Box Adaptive Network for Visual Tracking (CVPR 2020 · similarity 0.90)
- SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking (CVPR 2020 · similarity 0.89)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.87)
- DiffMOT: A Real-time Diffusion-based Multiple Object Tracker with Non-linear Prediction (CVPR 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.