Multi-Camera Multi-Human Multi-Task ($M^3$) Video Analysis


Direction 1: Air-ground-view multi-human video analysis

Data Hightlight : Air-ground-view human detection, association and tracking (IJCV 2024)

Project page: https://github.com/RuizeHan/CVMHATB

Method Hightlight : A new spatial distribution feature for human association (TPAMI 2022)

Project page: https://github.com/RuizeHan/CVMHAT

Other basics/extensions:

1. Deep spatial-distribution feature for human cross-view connection (CVPR 2022)

Project page: https://github.com/RuizeHan/DMHA

2. Human shadow based complementary-view camera pose estimation (IJCV 2023)

Project page: https://github.com/RuizeHan/CVCR

3. Complementary-view co-interest person detection (ACM MM 2020)

Project page: https://github.com/RuizeHan/CIP

example


Direction 2: Multi-view multi-human video analysis

Hightlight : Self-supervided spatial-temporal human association (ACM MM 2021, Oral, Submitted to TPAMI)

Project page: https://github.com/realgump/MvMHAT

Other basics/extensions:

1. Image based mului-view human association (TIP 2022)

Project page: https://github.com/RuizeHan/DAN4Ass

2. Mului-view human registration in birds-eye view (CVPR 2024)

Project page: https://github.com/zekunqian/BEVSee

example


Direction 3: Multi-human activity analysis

Hightlight : Panoramic human activity recognition (ECCV 2022)

Project page: https://github.com/RuizeHan/PAR

Other basics/extensions:

1. Human group detection (ECCV 2022, Oral)

Project page: https://github.com/Jiaoma/SHGD

2. Multi-view human interaction detection (ACM MM 2020)

Project page: https://github.com/RuizeHan/CVID

example