APS: a 3D human body posture set as a baseline for posture guidance

Human body postures are an important input modality for motion guidance and other application domains in HCI, e.g. games, character animations, and interaction with public displays. However, for training and guidance of body postures prior research had to define their own whole body gesture sets. Hence, the interaction designs and evaluation results are difficult to compare, due to a lack of a standardized posture set. In this work, we contribute APS (APS Posture Set), a novel posture set including 40 body postures. It is based on prior research, sports, and body language. For each identified posture, we collected 3D posture data using a Microsoft Kinect. We make the skeleton data, 3D mesh objects and SMPL data available for future research. Taken together, APS can be used to facilitate design of interfaces that use body gestures and as a reference set for future user studies and system evaluations.