Challenge Winners 2024
Track 1:
Winner: Team79 SJTU-LENOVO
Runner-up: Team221 YACHIYO
Overlap Suppression Clustering for Offline Multi-Camera People Tracking
The following table shows the top teams from the public leader board of Track 1 with paper submissions by the challenge submission deadline.
| Rank | Team ID | Team Name | Score(HOTA) | Online |
| 1 | 221 | Yachiyo | 71.9446 | No |
| 2 | 79 | SJTU-Lenovo | 67.2175 | Yes |
| 3 | 40 | Nota | 60.9261 | Yes |
| 4 | 142 | Fraunhofer IOSB | 60.8792 | Yes |
| 5 | 8 | UW-ETRI | 57.1445 | Yes |
| 6 | 50 | ARV | 51.0556 | Yes |
| 9 | 162 | Asilla | 40.3361 | Yes |
Track 2:
Winner: Team208 AliOpenTrek
CityLLaVA: Efficient Fine-Tuning for VLMs in City Scenario
Runner-up: Team28 AIO_ISC
The following table shows the top teams from the public leader board of Track 2 with paper submissions by the challenge submission deadline.
| Rank | Team ID | Team Name | Score (4 metrics avg.) |
| 1 | 208 | AliOpenTrek | 33.4308 |
| 2 | 28 | AIO_ISC | 32.8877 |
| 3 | 68 | Lighthouse | 32.3006 |
| 6 | 219 | UCF-SST-NLP | 29.0084 |
| 9 | 91 | HCMUS_AGAIN | 22.7371 |
Track 3:
Winner: Team155 TeleAI
Augmented Self-Mask Attention Transformer for Naturalistic Driving Action Recognition
Runner-up: Team5 SKKU-AutoLab
The following chart shows the performance from top teams on dataset A2 by the challenge submission deadline (for comparision) as well as on dataset B.
- As detailed in the track description, the award winners will be determined based on the performance on dataset B.
| Final Rank | Team ID | Team Name | Score on dataset A2 (os) | Score on dataset B (os) |
| 1 | 155 | TeleAI | 0.8282 | 0.8084 |
| 2 | 5 | SKKU-AutoLab | 0.7798 | 0.7187 |
| 3 | 165 | MCPRL | 0.6080 | 0.5978 |
Track 4:
Winner: Team9 VNPT AI
Robust Data Augmentation and Ensemble Method for Object Detection in Fisheye Camera Images
Runner-up: Team40 Nota
Road Object Detection Robust to Distorted Objects at the Edge Regions of Images
Honorable Mention: Team5 SKKU-AutoLab
Improving Object Detection to Fisheye Cameras with Open-Vocabulary Pseudo-Label Approach
The following table shows the top teams from the public leader board of Track 4 with paper submissions by the challenge submission deadline.
- We reproduced the results of the top 3 teams for a detailed and fair comparison and the rank of the top 3 remains unchanged. It is a very close run between team40 and team5. Team40 ends up winning the award as the runner-up and we do want to give team5 an honorable mention.
| Rank | Team ID | Team Name | Score (F1) |
| 1 | 9 | VNPT AI | 0.6406 |
| 2 | 40 | Nota | 0.6196 |
| 3 | 5 | SKKU-AutoLab | 0.6194 |
| 4 | 63 | UIT_AICLUB | 0.6077 |
| 5 | 15 | SKKU-NDSU | 0.5965 |
| 6 | 33 | MCPRL | 0.5883 |
Track 5:
Winner: Team99 UIT
Runner-up: Team76 China Mobile
An Effective Method for Detecting Violation of Helmet Rule for Motorcyclists
The following table shows the top teams from the public leader board of Track 5 with paper submissions by the challenge submission deadline.
| Rank | Team ID | Team Name | Score (mAP) |
| 1 | 99 | UIT | 0.4860 |
| 2 | 76 | China Mobile | 0.4824 |
| 3 | 9 | VNPT | 0.4792 |
| 7 | 57 | BUPT | 0.394 |