Abstract
In this contribution we present an algorithm for 2D-3D pose estimation of human beings. A human torso is modeled in terms of free-form surface patches which are extended with joints inside the surface. We determine the 3D pose and the angles of the arm joints from image silhouettes of the torso. This silhouette based approach towards human motion estimation is illustrated by experimental results for monocular or stereo image sequences. © Springer-Verlag 2004.
Cite
CITATION STYLE
Rosenhahn, B., Klette, R., & Sommer, G. (2004). Silhouette based human motion estimation. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3175, 294–301. https://doi.org/10.1007/978-3-540-28649-3_36
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
