Immense opportunity exists to make transportation systems smarter, based on sensor data from traffic, signaling systems, infrastructure, and transit.  Unfortunately, progress has been limited for several reasons — among them, poor data quality, missing data labels, and the lack of high-quality models that can convert the data into actionable insights There is also a need for platforms that can handle analysis from the edge to the cloud, which will accelerate the development and deployment of these models.

We are organizing the AI City Challenge Workshop at CVPR 2019 to help address these challenges by encouraging research and development into techniques that rely less on supervised approaches and more on transfer learning, unsupervised and semi-supervised approaches that go beyond bounding boxes.  It will focus on Intelligent Transportation System (ITS) problems, such as:

  • City-scale multi-camera vehicle tracking
  • City-scale multi-camera vehicle re-identification
  • Traffic anomaly detection – Leveraging unsupervised learning to detect anomalies such as lane violation, illegal U-turns, wrong-direction driving, etc.

We solicit original contributions in these and related areas where computer vision and specifically deep learning have shown promise in achieving large-scale practical deployment that will help make cities smarter.

To accelerate the research and development of techniques that rely less on supervised approaches and more on transfer learning, self-supervised and semi-supervised learning we are organizing this Challenge.


The datasets for the 2019 AI City Challenge, i.e., CityFlow (for MTMC vehicle tracking and re-identification) and the Iowa DOT Traffic Dataset (for traffic anomaly detection), can be accessed now. The evaluation server is open again for submission of test results. Please follow the Datasets Access Instructions to submit the completed Datasets Request Form to us. 



AI City Challenge CVPR Workshop 2019

09:00 – 09:30

Workshop Kickoff and Opening Comments

09:30 – 10:10

Two Oral Presentations (20 minutes each)

(1) Traffic Anomaly Detection via Perspective Map Based on Spatial-temporal Information Matrix
Shuai Bai; Zhiqun He; Yu Lei; Wei Wu; Chengkai Zhu; Ming Sun; Junjie Yan
(2) Anomaly Candidate Identification and Starting Time Estimation of Vehicles from Traffic Videos
Gaoang Wang; Xinyu Yuan; Aotian Zheng; Hung-Min Hsu; Jenq-Neng Hwang 


10:10 – 10:30

Coffee Break

10:30 – 12:10

Five Oral Presentations (20 minutes each)

(3) AI City Challenge 2019 – City-scale Video Analytics for Smart Transportation
Ming-Ching Chang; Jiayi Wei; Zheng-An Zhu; Yan-Ming Chen; Chan-Shuo Hu; Ming-Xiu Jiang; Chen-Kuo Chiang 
(4) Unsupervised Traffic Anomaly Detection Using Trajectories
Jianfei Zhao; Zitong Yi; Siyang Pan; Yanyun Zhao; Zhicheng Zhao; Fei Su; Bojin Zhuang
(5) Vehicle Re-identification with Learned Representation and Spatial Verification and Abnormality Detection with Multi-adaptive Vehicle Detectors for Traffic Video Analysis
Khac-Tuan Nguyen; Trung-Hieu Hoang; Minh-Triet Tran; Trung-Nghia Le; Ngoc-Minh Bui; Trong-Le Do; Viet-Khoa Vo-Ho; Quoc-An Luong; Mai-Khiem Tran; Thanh-An Nguyen; Thanh-Dat Truong; Vinh-Tiep Nguyen; Minh Do 
(6) A Comparative Study of Faster R-CNN Models for Anomaly Detection in 2019 AI City Challenge
Linu Shine; Anitha Edison; Jiji C.V. 
(7) Attention Driven Vehicle Re-identification and Unsupervised Anomaly Detection for Traffic Understanding
Pirazh Khorramshahi; Neehar Peri; Amit Kumar; Anshul Shah; Rama Chellappa


12:10 – 13:40


13:40 – 15:20

Five Oral Presentations (20 minutes each)

(8) Multi-camera Vehicle Tracking and Re-identification Based on Visual and Spatial-temporal Features
Xiao Tan; Zhigang Wang; Minyue Jiang; Xipeng Yang; Jian Wang; Yuan Gao; Xiangbo Su; Xiaoqing Ye; Yuchen Yuan; Dongliang He; Shilei Wen; Errui Ding 
(9) Multi-view Vehicle Re-identification Using Temporal Attention Model and Metadata Re-ranking
Tsung-Wei Huang; Jiarui Cai; Hao Yang; Hung-Min Hsu; Jenq-Neng Hwang 
(10) Partition and Reunion: A Two-branch Neural Network for Vehicle Re-identification
Hao Chen; Benoit Lagadec; Francois Bremond
(11) Multi-camera Vehicle Tracking with Powerful Visual Features and Spatial-temporal Cue
Zhiqun He; Yu Lei; Shuai Bai; Wei Wu
(12) Vehicle Re-identifiation and Multi-camera Tracking in Challenging City-scale Environment
Jakub Špaňhel; Vojtěch Bartl; Roman Juránek; Adam Herout

15:20 – 15:50

Coffee Break

15:50 – 16:30

Two Oral Presentations (20 minutes each)

(13) Multi-camera Tracking of Vehicles based on Deep Features Re-ID and Trajectory-based Camera Link Models
Hung-Min Hsu; Tsung-Wei Huang; Gaoang Wang; Jiarui Cai; Zhichao Lei; Jenq-Neng Hwang 
(14) A Locality Aware City-scale Multi-Camera Vehicle Tracking System
Yunzhong Hou; Heming Du; Liang Zheng 


16:30 – 17:00

Announcement of Challenge Winners and Awards Ceremony

17:00 – 18:00

Poster/Demo Session




Milind Naphade

NVIDIA Corporation

Rama Chellappa

University of Maryland, College Park

David Anastasiu

San Jose State University

Anuj Sharma

Iowa State University

Ming-Ching Chang

University at Albany – SUNY

Ming-Yu Liu

NVIDIA Research

Xiaodong Yang

NVIDIA Research

Siwei Lyu

University at Albany – SUNY

Jenq-Neng Hwang

University of Washington, Seattle


Our paper based on the benchmarks of Track 1 and Track 2 of the 2019 AI City Challenge has been accepted for oral presentation at the CVPR 2019 main conference. The teams participating in the 2019 AI City Challenge will be among the first researchers working on this benchmark, and we believe your work will create a significant impact in cross-camera vehicle tracking and re-identification. Please note that the official name of our benchmark for Track 1 is “CityFlow,” and the benchmark of Track 2 is a subset of that, named “CityFlow-ReID.” The paper is available here. The reference to our paper is posted below. 

author = {Tang, Zheng and Naphade, Milind and Liu, Ming-Yu and Yang, Xiaodong and Birchfield, Stan and Wang, Shuo and Kumar, Ratnesh and Anastasiu, David and Hwang, Jenq-Neng},
title = {CityFlow: A City-Scale Benchmark for Multi-Target Multi-Camera Vehicle Tracking and Re-Identification},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2019}

The Challenge summary paper (The 2019 AI City Challenge) has been published in the proceedings of 2019 CVPRW. Please refer to these two papers when you continue your work on this dataset.

author = {Naphade, Milind and Tang, Zheng and Chang, Ming-Ching and Anastasiu, David C. and Sharma, Anuj and Chellappa, Rama and Wang, Shuo and Chakraborty, Pranamesh and Huang, Tingting and Hwang, Jenq-Neng and Lyu, Siwei},
title = {The 2019 AI City Challenge},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month = {June},
year = {2019},
pages = {452–460}