2022 WORKSHOP PROGRAM

This year’s CVPR main conference will be hosting at New Orleans, Louisiana, US. Our AI City Challenge workshop this year will be mostly on-site and combined with virtual attendance support.

Note: All times are based on US Central Time Zone (CDT).

9:00 AM – 9:30 AM

Opening – Workshop Organization Presentation (recording)

9:30 AM – 10:15 AM

Keynote

3D Computer Visison for Dynamic Scene Understanding (presentation recording)

Speaker: Daniel Cremers, Technische Universität München

10:15 AM – 10:20 AM

Break (5 minutes)

10:20 AM – 11:20 AM

 

Paper Presentations – Track1 (60 minutes, 45 mins pre-recorded videos)

(1) Box-Grained Reranking Matching for Multi-Camera Multi-Target Tracking – 10 minutes (media)

Yang, Xipeng; Ye, Jin*; Lu, Jincheng; Gong, Chenting; Jiang, Minyue; Lin, Xiangru; Zhang, Wei; Tan, Xiao; Li, Yingying; Ye, Xiaoqing; Ding, Errui

(2) Multi-Camera Vehicle Tracking System for AI City Challenge 2022 – 10 minutes (media)

Nie, Ding*; Li, Fei; Wang, Zhen

(3) City-Scale Multi-Camera Vehicle Tracking based on Space-Time-Appearance Features – 5 minutes (media)

Yao, Hui*; Duan, Zhizhao; Xie, Zhen; Chen, Jinbo; Wu, Xi; Xu, Duo; Gao, Yutao

(4) Improving Multi-Target Multi-Camera Tracking by Track Refinement and Completion – 5 minutes (media)

Specker, Andreas*; Florin, Lucas; Cormier, Mickael; Beyerer, Jürgen

(5) A Robust Traffic-Aware City-Scale Multi-Camera Vehicle Tracking Of Vehicles – 5 minutes (media)

Nguyen-Ngoc Tran, Duong*; Pham, Long H; Jeon, Hyung-Joon; Nguyen, Huy Hung; Jeon, Hyung-Min; Huu-Phuong Tran, Tai; Jeon, Jae

(6) Multi-Camera Multi-Vehicle Tracking with Domain Generalization and Contextual Constraints – 5 minutes (media)

Chung, Nhat Minh; Le, Huy; Nguyen, Vuong; Nguyen, Quang; Nguyen, Thong; Thai, Tin; Ha, Synh Viet-Uyen*

(7) Multi-Camera Vehicle Tracking Based on Occlusion-aware and Inter-vehicle Information – 5 minutes (media)

Liu, Yuming*; Zhang, Bingzhen; Zhang, Xiaoyong; Wang, Sen; Xu, Jianrong

11:20 AM – 12:10 PM

Paper Presentations – Track2 (50 minutes, 40 mins pre-recorded videos)

(1) A Multi-granularity Retrieval System for Natural Language-based Vehicle Retrieval – 10 minutes (media)

Lin, Xiangru*; Zhang, Jiacheng; Jiang, Minyue; Yu, Yue; Gong, Chenting; Zhang, Wei; Tan, Xiao; Li, Yingying; Ding, Errui; Li, Guanbin

(2) Tracked-Vehicle Retrieval by Natural Language Descriptions With Domain Adaptive Knowledge – 10 minutes (media)

Le, Huy; Nguyen, Quang; Nguyen, Vuong; Nguyen, Thong; Chung, Nhat Minh; Thai Trung, Tin; Ha, Synh Viet-Uyen*

(3) Symmetric Network with Spatial Relationship Modeling for Natural Language-based Vehicle Retrieval – 5 minutes (media)

Zhao, Chuyang*; chen, haobo; Zhang, Wenyuan; Chen, Junru; Sipeng, Zhang; li, yadong; Li, Boxun

(4) Text Query based Traffic Video Event Retrieval with Global-Local Fusion Embedding – 5 minutes (media)

Pham, Khoi Minh; Nguyen-Ho, Thang-Long; Nguyen, Tien-Phat; Do, Minh; Nguyen, Tam; Tran, Minh-Triet*

(5) Natural Language-Based Vehicle Retrieval with Explicit Cross-Modal Representation Learning – 5 minutes (media)

Xu, Bocheng; Yihua, Xiong; Zhang, Rui*; Feng, Yanyi; Wu, Haifeng

(6) OMG: Observe Multiple Granularities for Natural Language-Based Vehicle Retrieval – 5 minutes (media)

Du, Yunhao; Zhang, Binyu; RUAN, XIANGNING; Su, Fei; Zhao, Zhicheng*; Chen, Hong

12:10 PM – 13:00 PM

Lunch Break (50 minutes)

01:00 PM – 02:20 PM

Paper Presentations – Track 3 (80 mins, 60 mins pre-recorded videos)

(1) An Effective Temporal Localization Method with Multi-View 3D Action Recognition for Untrimmed Naturalistic Driving Videos – 10 minutes (media)

Bui, Nam Khac Hoai*

(2) Stargazer: A Transformer-based Driver Action Detection System for Intelligent Transportation – 10 minutes (media)

Liang, Junwei*; Zhu, He; Zhang, Enwei; Zhang, Jun

(3) Learning Generalized Feature for Temporal Action Detection: Application for Natural Driving Action Recognition Challenge – 5 minutes (media)

Nguyen, Chuong H*

(4) MV-TAL: Mulit-view Temporal Action Localization in Naturalistic Driving – 5 minutes (media)

Li, Wei*; Chen, Shimin; Gu, Jianyang; Wang, Ning; Chen, Chen; Guo, Yandong

(5) Federated Learning-based Driver Activity Recognition for Edge Devices – 5 minutes (media)

Doshi, Keval*; Yilmaz, Yasin

(6) PAND: Precise Action Recognition on Naturalistic Driving – 5 minutes (media)

zhao, Hangyue; Xiao, Yuchao; Zhao, Yanyun *

(7) A Coarse-to-Fine Boundary Localization method for Naturalistic Driving Action Recognition – 5 minutes (media)

Chen, Zhenzhong*; Ding, Guanchen; Han, Wenwei; Wang, Chenglong; Cui, MingPeng; Zhou, Lin; Pan, Dianbo; Wang, Jiayi; Zhang, Junxi

(8) Density-Guided Label Smoothing for Temporal Localization of Driving Actions – 5 minutes (media)

Alkanat, Tunc; Akdag, Erkut*; Bondarev, Egor; de With, P. H. N.

(9) Temporal Driver Action Recognition using Action Classification Method – 5 minutes (media)

Alyahya, Munirah N*; ALhussan, Taghreed M; Alghannam, Shahad

(10) Key Point-Based Driver Activity Recognition – 5 minutes (media)

Vats, Arpita; Anastasiu, David C*

02:20 PM – 03:05 PM

Paper Presentations – Track 4 (45 mins, 35 mins pre-recorded videos)

(1) Amazing Results With Limited Data In Multi-Class Product Counting and Recegnition – 10 minutes (media)

Wan, Junfeng*; Shuhao, Qian; Tian, Zihan; Zhao, Yanyun

(2) DeepACO: A Robust Deep Learning-based Automatic Checkout System – 10 minutes (media)

Pham, Long H*; Nguyen-Ngoc Tran, Duong; Nguyen, Huy Hung; Huu-Phuong Tran, Tai; Jeon, Hyung-Joon; Jeon, Hyung-Min; Jeon, Jae

(3) VISTA: Vision Transformer enhanced by U-Net and Image Colorfulness Frame Filtration for Automatic Retail Checkout – 5 minutes (media)

Shihab, Md. Istiak Hossain; Tasnim, Nazia*; Zunair, Hasib; Rupty, Labiba K; Mohammed, Nabeel

(4) A Region-Based Deep Learning Approach to Automated Retail Checkout – 5 minutes (media)

Shoman, Maged*; Aboah, Armstrong; Morehead, Alex; Duan, Ye; Daud, Abdulateef; Adu-Gyamfi, Yaw

(5) PersonGONE: Image Inpainting for Automated Checkout Solution – 5 minutes (media)

Bartl, Vojtěch*; Špaňhel, Jakub; Herout, Adam

03:05 PM – 03:10 PM

Paper Presentations – Additional (5 mins, 5 mins pre-recorded videos)

(From Track1 of the 2021 Challenge)

(1) Detecting Vehicles on the Edge: Knowledge Distillation to Improve Performance in Heterogeneous Road Traffic – 5 minutes (media)

Bharadhwaj, Manoj*; Ramadurai, Gitakrishnan; Ravindran, Balaraman

03:10 PM – 03:25 PM

 

Break (15 minutes)

03:25 PM – 03:45 PM

Open Discussion and Summary of Challenges (20 minutes)

03:45 PM – 04:45 PM

 

Panel Discussion (recording)

panelists: Milind Naphade, Ming-Ching Chang, David Anastasiu, Anuj Sharma, Liang Zheng

Topic1: Ethics and Privacy of AI for Smart City and Social Good (30 mins)

Topic2: Synthetic Data Generation for Large-Scale Training and Deployment of AI Models (30 mins)

04:45 PM – 05:00 PM

 

Award Ceremony (recording)

05:00 PM

 

Adjourn