Research map / Detection and segmentation / Tracking
CityFlow: A City-Scale Benchmark for Multi-Target Multi-Camera Vehicle Tracking and 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 probabilistic camera-transition modeling with FVS rather than PROVID re-ranking | TC + BA MTMC tracking · CityFlow MTMC | IDF1: 40.6% -> 46.3% | Helped |
| Replace PROVID with online-learned transition modeling (2WGMMF) | TC + BS MTMC tracking · CityFlow MTMC | IDF1: 39.0% -> 45.1% | Helped |
| Add cross-entropy loss to hard triplet loss | DenseNet121 person ReID · CityFlow-ReID | mAP: 30.5 -> 31.0 | Helped |
| Replace hard triplet loss with center loss | DenseNet121 person ReID · CityFlow-ReID | mAP: 30.5 -> 10.7 | Hurt |
| Replace batch-all sampling with batch-hard sampling | MobileNetV1 vehicle ReID · CityFlow-ReID | mAP: 31.3% -> 32.0% | Helped |
| Replace Bhattacharyya distance with L∞ distance | FVS vehicle ReID · CityFlow-ReID | mAP: 6.3% -> 3.2% | Hurt |
| Use SSD512 instead of YOLOv3 for detection | TC single-camera tracking · CityFlow MTSC | MOTA: 68.0% -> 70.3% | Helped |
| Use Faster R-CNN instead of SSD512 for detection | TC single-camera tracking · CityFlow MTSC | MOTA: 70.3% -> 68.4% | Hurt |
| Use TC instead of DeepSORT for single-camera tracking | SSD detection · CityFlow MTSC | IDF1: 79.5% -> 79.7% | No clear effect |
| Use FVS rather than PROVID spatio-temporal association | TC + Xent MTMC tracking · CityFlow MTMC | IDF1: 35.2% -> 40.5% | Helped |
| Change vehicle ReID sampling from batch-all to batch-weighted | MobileNetV1 vehicle ReID · CityFlow-ReID | rank-1: 49.6% -> 50.1% | Mixed |
Closest papers by meaning
- On-Road Object Importance Estimation: A New Dataset and A Model with Multi-Fold Top-Down Guidance (NeurIPS 2024 · similarity 0.85)
- UrbanIng-V2X: A Large-Scale Multi-Vehicle, Multi-Infrastructure Dataset Across Multiple Intersections for Cooperative Perception (NeurIPS 2025 · similarity 0.85)
- MTMMC: A Large-Scale Real-World Multi-Modal Camera Tracking Benchmark (CVPR 2024 · similarity 0.85)
- Traffic Scene Parsing through the TSP6K Dataset (CVPR 2024 · similarity 0.85)
- Bridging Traffic State and Trajectory for Dynamic Road Network and Trajectory Representation Learning (AAAI 2025 · similarity 0.84)
- eTraM: Event-based Traffic Monitoring Dataset (CVPR 2024 · similarity 0.83)
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.