Alternative strategies for runtime facial motion capture

Abstract

classroom use is granted without fee provided that copies are not made or distributed for commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the Owner/Author. Facial motion capture has been hitherto, an effective albeit costly means of delivering performances for game characters. Using Kinect hardware we consider a pipeline for delivering game ready performances in Unity, enlisting the talents of actors and game developers. We begin with a review of the following data acquisition pipeline as a basis for motion capture: This data acquisition process informs a pipeline based on the concept of skeletal retargeting whereby the motion capture data stream may be mapped back to a common joint based skeletal system thus rendering it scalable and friendly for implementation into the game development pipeline. In this presentation we hope to take a look at the facial motion capture data set and its applicability to varied characters. Aside from the technical challenges of creating assets for facial motion capture, there is however, the problem of credibly reproducing performances of subjects. As the closer one moves towards realism, the harder it becomes to create an empathic human face. With this in mind we address the phenomenon commonly described as the " uncanny valley " in relation to motion captured facial performances and attempt to define the limits of the technology. 3 Conclusion Accessible runtime facial motion capture is an area of growing interest. The advent of the Kinect as a PC peripheral and Mi-crosoft " s Kinect Fusion Project give us a glimpse into the possibilities afforded by this nascent technology. The implications of such technology are wide-ranging and ultimately offer the prospect of revolutionizing interactive entertainment. Sequence of expressions with corresponding facial reference.