Inertial Poser

Inertial Poser research focuses on accurately estimating 3D human body pose and movement using data from sparsely placed inertial measurement units (IMUs), offering a less intrusive alternative to traditional motion capture systems. Current research emphasizes improving accuracy by fusing IMU data with other sensor modalities like ultra-wideband ranging or LiDAR, and employing advanced algorithms such as graph-based machine learning models, part-based dynamic models, and transformer networks to address challenges like drift and jitter. These advancements are significant because they enable more affordable and convenient motion capture for applications ranging from animation and virtual reality to healthcare and sports analysis.