Workshop on 

New Challenges in 3D Human Understanding


@CVPR 2024, Seattle, USA

June 17th, 2024

Overview

Understanding humans in 3D remains at the heart of computer vision advancements. The complexities associated with analyzing human behaviors, postures, and interactions are multifaceted, making them a challenging yet crucial aspect of vision research. Recent developments in the field have brought to light new challenges, including multi-human interaction analysis, fine-grained motion capture, clothed body reconstruction, virtual try-ons and tracking under extreme occlusions. This workshop aims to gather researchers in this field, to assess existing methodologies, understand their limitations, and collaboratively discuss novel approaches to push the boundaries of 3D human understanding in computer vision. To that end, this workshop will feature: (1) Invited speakers from both the academic and industrial spheres. (2) An invited poster session that covers diverse relevant works accepted at recent top-tier conferences such as CVPR 2024 and NeurIPS/ICCV/CVPR 2023. (3) A panel discussion featuring invited speakers, aimed at fostering the exchange of ideas, addressing limitations, and identifying promising research directions.


Human understanding with various scenarios and challenges...

During social interactions

Under occlusions

With fine-grained motions

In the crowd

From egocentric views

In 3D environments

Speakers

Michael Black

MPI for Intelligent Systems

C. Karen Liu

Stanford

Jitendra Malik

UC Berkeley

Organizers

Qianli Ma

NVIDIA

Siwei Zhang

ETH Zurich

Shashank Tripathi

MPI for Intelligent Systems

Senior Advisory Board

Angjoo Kanazawa

UC Berkeley

Siyu Tang

ETH Zurich