RegisterGet new password
ISMAR 2014 - Sep 10-12 - Munich, Germany

ieee.org
computer.org
vgtc.org

ISMAR Sessions

Call for Participation

Committees

Supporters and Media

Contact




Social Media

LinkedInLinkedIn



Previous Years

Measurements of Live Actor Motion in Mixed Reality Interaction

SCHEDULE INFORMATION

Event TitleSession TitleChairRoomStartEnd
User InterfacesUser InterfacesSteven Feiner, Columbia UniversityHS111 Sep, 2014 02:00 PM11 Sep, 2014 03:45 PM
Authors: 
Gregory Hough
Authors: 
Ian Williams
Authors: 
Cham Athwal
Abstract: 
This paper presents a method for measuring the magnitude and impact of errors in mixed reality interactions. We define the errors as measurements of hand placement accuracy and consistency within bimanual movement of an interactive virtual object. First, a study is presented which illustrates the amount of variability between the hands and the mean distance of the hands from the surfaces of a common virtual object. The results allow a discussion of the most significant factors which should be considered in the frame of developing realistic mixed reality interaction systems. The degree of error was found to be independent of interaction speed, whilst the size of virtual object and the position of the hands are significant. Second, a further study illustrates how perceptible these errors are to a third person viewer of the interaction (e.g. an audience member). We found that interaction errors arising from the overestimation of an object surface affected the visual credibility for the viewer considerably more than an underestimation of the object. This work is presented within the application of a real-time Interactive Virtual Television Studio, which offers convincing real-time interaction for live TV production. We believe the results and methodology presented here could also be applied for designing, implementing and assessing interaction quality in many other Mixed Reality applications.

Sponsors (Become one)

Diamond



Platinum



Gold




Bronze






SME







Academic