Close

Presentation

Fast and Scalable Human Pose Estimation using mmWave Point Cloud
TimeWednesday, July 13th4:30pm - 4:50pm PDT
Location3000, Level 3
Event Type
Research Manuscript
Keywords
Design of Cyber-physical Systems, Cloud Computing and IoT
Topics
Cloud
Design
DescriptionHuman pose estimation offers valuable insight for autonomous driving and health monitoring due to its accurate visual representation for human motions. The emerging millimeter-wave (mmWave) device can generate a high-quality 3D point cloud for human pose estimation. However, the point cloud from mmWave is sparser compared to other sources. In addition, the scarce mmWave datasets make machine learning (ML) algorithms harder to adapt to new users and movements since these algorithms need vast amounts of data.