Computer Science

Localization and recognition of human action in 3D using transformers

J. Sun, L. Huang, et al.

Discover the groundbreaking BABEL-TAL dataset and the LocATe model that sets new benchmarks in 3D action localization! This innovative research, conducted by Jiankai Sun and colleagues, pushes the boundaries of 3D human behavior analysis with applications in HCI and healthcare.

00:00

Playback language: English

Index

Abstract

This paper introduces BABEL-TAL (BT), a new benchmark dataset for 3D action localization, addressing limitations in existing datasets. A novel model, LocATe (Localizing Actions with Transformers), is proposed, achieving state-of-the-art performance on BT and PKU-MMD, even with limited training data. The research advances 3D human behavior analysis with applications in HCI and healthcare.

Publisher

Communications Engineering

Published On

Sep 03, 2024

Authors

Jiankai Sun, Linjiang Huang, Hongsong Wang, Chuanyang Zheng, Jianing Qiu, Md Tauhidul Islam, Enze Xie, Bolei Zhou, Lei Xing, Arjun Chandrasekaran, Michael J. Black

Related Publications

Explore these studies to deepen your understanding of the subject.

Engineering and Technology

Application of 3D-printed tissue-engineered skin substitute using innovative biomaterial loaded with human adipose-derived stem cells in wound healing

H. Fu, D. Zhang, et al.

Medicine and Health

Automated detection of intracranial aneurysms using skeleton-based 3D patches, semantic segmentation, and auxiliary classification for overcoming data imbalance in brain TOF-MRA

S. Ham, J. Seo, et al.

Engineering and Technology

Using visible light to activate antiviral and antimicrobial properties of TiO2 nanoparticles in paints and coatings: focus on new developments for frequent-touch surfaces in hospitals

M. Schutte-smith, E. Erasmus, et al.

Interdisciplinary Studies

The political and social contradictions of the human and online environment in the context of artificial intelligence applications

R. Rakowski and P. Kowaliková

Listen, Learn & Level Up

Over 10,000 hours of research content in 25+ fields, available in 12+ languages.

No more digging through PDFs, just hit play and absorb the world's latest research in your language, on your time.

listen to research audio papers with researchbunny