“Shall We Dance?” by Ebihara

  • ©Kazuyuki Ebihara


Entry Number: 26


    Shall We Dance?

Program Title:

    Enhanced Realities



Project Affiliation:

    ATR Media Integration and Communications Research Lab and University of Maryland


    Real-time 3D computer vision gives users control over both the movement and facial expression of a virtual puppet and the music to which the puppet “dances.” Multiple cameras observe a person, and human silhouette analysis achieves real-time 3D estimation of human postures. Facial expressions are estimated from images acquired by a viewing-direction controllable camera, so that the face can be tracked. From the facial images, deformations of each facial component are estimated. The estimated body postures and facial expressions are reproduced in the pup- pet model by deforming the model according to the estimated data. All the estimation and rendering processes run in real time on PC-based systems. Attendees can see themselves dancing in a virtual scene as virtual puppets.

Additional Images:

©Kazuyuki Ebihara ©Kazuyuki Ebihara ©Kazuyuki Ebihara ©Kazuyuki Ebihara


Overview Page: