Challenge Winners

Track 1:

Winner: Baidu

Robust Movement-Specific Vehicle Counting at Crowded Intersections

Runner-up: DiDi & Huaqiao University

Robust and Fast Vehicle Turn-counts at Intersections via an Integrated Solution from Detection, Tracking and Trajectory Modeling

 

The following table shows the performance of submitted codes from top teams on Dataset B (4.5 hours). All submitted codes were tested on the same Ubuntu machine with 32GB RAM, i7-5930K CPU (12 Cores) and one NVIDIA TITAN Xp GPU (12GB memory).

RankTeam ID OrganizationRun Time Run SpeedEffectivenessEfficiencyFinal Score
199Baidu3.24 hours14.31 fps0.9460.8540.918
260DiDi – MapVision2.67 hours17.35 fps0.9050.880.898
326DiDi – orange control1.33 hours34.82 fps0.8750.940.894
4110ENGIE6.92 hours6.71 fps0.9460.6890.869
592Carnegie Mellon University7.22 hours6.42 fps0.9040.6750.836
680Ho Chi Minh City University of Science1.39 hours33.35 fps0.3240.9370.508

Track 2:

Winner: Baidu & University of Technology Sydney

Going Beyond Real Data: A Robust Visual Representation for Vehicle Re-identification

Runner-up: RuiYan Technology

VOC-ReID: Vehicle Re-identification based on Vehicle-Orientation-Camera

 

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

RankTeam IDTeam NameScore
173Baidu-UTS0.8413
242RuiYanAI0.7810
339DMT0.7322
436IOSB-VeRi0.6899
530BestImage0.6684
644BeBetter0.6683
772UMD_RC0.6668
87Ainnovation0.6561
946NMB0.6206
1081Shahe0.6191

Track 3:

Winner: Carnegie Mellon University

ELECTRICITY: An Efficient Multi-camera Vehicle Tracking System for Intelligent City

Runner-up: Xi’an Jiaotong University

City-Scale Multi-Camera Vehicle Tracking by Semantic Attribute Parsing and Cross-Camera Tracklet Matching

The following table shows the top teams from the public leader board of Track 3 by the challenge submission deadline.

RankTeam IDTeam NameScore
192INF0.4616
211XJTU-Alpha0.4400
363DukBaeGi0.3483
4111EINI_CQUPT0.3411
572UMD_RC0.1245
675Albany_NCCU0.0620
730Youtu0.0452
831SJTU_yutinggao0.0387

Track 4:

Winner: Baidu & Sun Yat-sen University

Multi-Granularity Tracking with Modularlized Components for Unsupervised Vehicles Anomaly Detection

Runner-up: University of South Florida

Fast Unsupervised Anomaly Detection in Traffic Videos

The following table shows the top teams from the public leader board of Track 4 by the challenge submission deadline.

RankTeam IDTeam NameScore
1113BaiduVis0.9695
251SIS Lab0.6625
3106CETCVLAB0.5438
472UMD_RC0.2952
591HappyLoner0.2909
626Orange-Control0.2386
749PapaNet0.1703
8132Team_Gaze_NSU_UAP0.0958