Institutional Repository
Technical University of Crete
EN  |  EL



My Space

Near-contact person-to-3D character dance training: comparing AR and VR for interactive entertainment

Kirakosian Salva, Daskalogrigorakis Grigorios, Maravelakis, Emmanuel, Mania Aikaterini

Full record

Year 2021
Type of Item Conference Publication
Bibliographic Citation S. Kirakosian, G. Daskalogrigorakis, E. Maravelakis and K. Mania, "Near-contact person-to-3D character dance training: comparing AR and VR for interactive entertainment," presented at the 2021 IEEE Conference on Games (CoG), Copenhagen, Denmark, 2021, doi: 10.1109/CoG52621.2021.9619037.
Appears in Collections


This paper presents an innovative, near-contact, person-to-3D character, real-time dance training framework for leaders developed in Augmented Reality (AR) for interactive entertainment. Rather than mimicking pre-recorded dance animations as in previous work, the user is trained to lead a 3D partner while dancing, employing mid-air gestures. Real and virtual hands are joined and dance movements are initiated to which the 3D character responds to, through Unity's Inverse Kinematics solvers, in a non-predetermined manner following Latin dance rules. We evaluate task workload and usability of our AR dance training system compared with the Virtual Reality (VR) equivalent. Perceived training performance, accuracy of actions and fun were lower in AR, evoking increased training effort in AR compared to playfully training in VR. Design recommendations include increasing AR's Field of View, enhancing high quality graphics in AR, removing obtrusive cabling in VR and improving finger tracking in VR.