Challenge Winners 2023

Track 1:

Winner: Team6 UW-ETRI

Enhancing Multi-Camera People Tracking with Anchor-Guided Clustering and Spatio-Temporal Consistency ID Re-Assignment

Runner-up: Team9 HCMIU

Multi-camera People Tracking With Mixture of Realistic and Synthetic Knowledge

 

The following table shows the top teams from the public leader board of Track 1 by the challenge submission deadline.

RankTeam ID Team NameScore(IDF1)
16UW-ETRI0.9536
29HCMIU0.9417
341SJTU-Lenovo0.9331
451Fraunhofer IOSB0.9284
5113HUST0.9207
1038Nota0.8676
1320SKKU0.6171

Track 2:

Winner: Team9 HCMIU

Tracked-Vehicle Retrieval by Natural Language Descriptions With Multi-Contextual Adaptive Knowledge

Runner-up: Team28 Lenovo

A Unified Multi-modal Structure for Retrieving Tracked Vehicles through Natural Language​ Descriptions

 

The following table shows the top teams from the public leader board of Track 2 by the challenge submission deadline.

RankTeam IDTeam NameScore
19HCMIU0.8263
228Lenovo0.8179
385HCMUS0.4795

Track 3:

Winner: Team209 Meituan

Multi View Action Recognition for Distracted Driver Behavior Localization

Runner-up: Team60 JNU

Action Probability Calibration for Efficient Naturalistic Driving Action Localization

The following chart shows the performance of submitted codes from top teams on test set B.

  • We could not finished evaluation on Team49’s submission because the code is still bugged after a couple iterations. Thus remove Team49 from the following evaluation result table.
Final RankTeam IDTeam NameScore on TestB (activity overlap score)
1209Meituan-IoTCV0.2298
260JNU0.1946
383Viettel0.1836
48Prude Digital Twin Lab0.1669

Track 4:

Winner: Team23 Fujitsu

CheckSORT: Refined Synthetic Data Combination and Optimized SORT for Automatic Retail Checkout

Runner-up: Team21 BUPT

AdaptCD: An Adaptive Target Region-based Commodity Detection System

The following chart shows the performance of submitted codes from top teams on test set B.

  • We could not finished evaluation on Team1’s submission because of code bugs and insufficient documentation. Thus remove Team1 from the following evaluation result table.
RankTeam IDTeam NamePrecisionRecallF1 Score
123Fujitsu R&D Center0.85120.83740.8443
221BUPT0.71720.84540.7761
333SKKU Automation Lab0.72030.69110.7054
413UToronto0.51590.65850.5786
5200Centific0.47670.66670.5559

 

Track 5:

Co-Winner: Team33 SKKU Automation Lab

Robust Automatic Motorcycle Helmet Violation Detection for an Intelligent Transportation System

Co-Winner: Team58 CTC-AI

An Effective Motorcycle Helmet Object Detection Framework for Intelligent Traffic Safety

The following table shows the top teams from the public leader board of Track 5 by the challenge submission deadline.

RankTeam ID Team NameScore(IDF1)
133SKKU Automation Lab0.4163
258CTC-AI0.4107
3111GRAPH@FIT0.3875
437SMARTVISION0.3682
5219GroundTruth0.3670
6220HCMUS0.3664
718UT_He0.3452
816UT_NYCU_SUNY-Albany0.3448
930BUPT_MCPRL0.3356
1036Helios0.3342