Zetong Yang

Chinese University of Hong Kong, Start Making A Reader Today

Papers

Total Citations

172

H-Index

About

Zetong Yang is a researcher specializing in 3D perception, autonomous driving, and robotics, with a particular focus on point cloud understanding and object detection. His work addresses some of the most pressing challenges in enabling machines to accurately interpret three-dimensional environments in real time. Yang's most notable contribution is the 3D Multi-frame Attention Network (3D-MAN), published in 2021, which garnered over 114 citations. This work tackled a critical limitation in existing 3D object detection methods — their reliance on single-frame data — by developing an attention-based architecture that effectively aggregates information across multiple temporal frames, significantly improving detection accuracy for autonomous driving and robotics applications. Building on this foundation, Yang co-authored "A Unified Query-based Paradigm for Point Cloud Understanding" (2022, 51 citations), introducing the Embedding-Querying (EQ-Paradigm), a versatile framework unifying detection, segmentation, and classification tasks within a single architecture. This contribution reflects a broader ambition to streamline and standardize 3D scene understanding pipelines. With a growing citation record and research that bridges theoretical innovation with real-world applicability, Zetong Yang is an emerging voice in the autonomous systems and 3D computer vision communities, producing work that continues to influence both academic research and industry development.