IEEE Consumer Electronics Magazine - July 2018 - 48

To try on their outfit, participants were instructed to walk
toward one of the two mirrors, stand directly in front of it,
and then press a button on their controller. Regardless of the
mirror they chose, the outcome was the same; their avatars
performed an animation, i.e., simulated the removal of their
pants. The original clothing then instantly disappeared, revealing a naked female or male VB with realistic, fully visible
genitalia (Figure 5).
After being left for a few seconds without any further
instructions, and while still wearing the HMD, participants
were asked to state their dominant emotion. This concluded the
experimental session and the participants were then instructed
to remove the equipment. (Footage of the participants and
screen recordings of the VR experience are available in [21].)

matErials anD ExPErimEntal sEtuP
To implement all of the previously mentioned experiment
conditions, a number of processes (P) were designed and
developed using several software and hardware tools. Cinema
4D was used for creating the necessary 3-D models (e.g.,
walls, mirrors), assigning textures, adjusting the lighting conditions, and finally, for exporting the models for use into the

FiGuRe 4. A virtual hand reaching to pick up the outfit.

FiGuRe 5. The naked male and female avatars (no censorship was
used for the experiment).
48 IEEE Consumer Electronics Magazine

July 2018

application's main engine, Unity 3D [Figure 6(b)]. Cinema
4D was also used to accomplish a task known as texture baking, a process in which the effects of light on static objects in
a 3-D scene are calculated and the results are written to textures that are overlaid on top of scene geometry to create the
effect of lighting.
The 3-D avatar models representing the female and male
users as well as the salesman were created with Adobe Fuse,
a character-design software that features customizable,
human 3-D models. All 3-D avatar models were then imported into Mixamo, a cloud-based platform where 3-D avatar
models can be rigged, skinned [i.e., the process of creating a
skeleton of joints (bones) to control the movements of a 3-D
character and then attaching the model mesh to the skeleton],
and assigned with a variety of prerecorded motion-capture
animations. Finally, the rigged, skinned, and animated 3-D
avatar models were imported into Unity 3D.
The female and male 3-D avatars were programmed to be
controlled (e.g., walk and interact) by the participants with
the use of a Logitech wireless controller, as demonstrated in
Figure 6. For group B, the salesman's 3-D avatar was programmed to perform a few animations (e.g., deliver an auditory and animated greeting when the participant walks into
the store).
As previously mentioned, the salesman for group C was
controlled by a real actor. This was achieved by recording
motion-capture data in real time with the Vicon Blade system, which consists of infrared cameras that record the movement of retroreflective markers placed in specific spots of a
suit. The captured data were retargeted into Vicon Pegasus
and the rigged 3-D salesman character, which was previously
imported into Unity 3D by using a Pegasus Unity 3D plugin.
Therefore, a real actor could move around a room, with his
movement redirected to the salesman 3-D character in real
time (Figure 6). (An example of this process may be viewed
in [22].) To allow the actor to be able to track the participants' location within the VE, a 40-in wide-screen monitor
was placed in the motion-capture room (i.e., the salesman's
room) displaying from top-down, the entire virtual clothing
store space in real time.
The last part of the VR application involved the delivery
of the VR world from Unity 3D to the participants' screen.
The main equipment used for this was a Shinecon VR HMD,
in which a Samsung Galaxy S5 Neo featuring a 5.1 full
high-definition screen was placed [Figure 6(a)]. Additionally, Trinus VR enabled the wireless transmission of the smartphone's orientation sensor data to Unity 3D to match the
Unity 3D camera's orientation with the participants' head
orientation in real time. Trinus VR also created a stereoscopic image based on the Unity 3D camera content, and wirelessly disseminated it to the smartphone placed in the HMD
so that the participant could properly view the VR content
(Figure 6). Finally, the researcher who controlled the salesman and all of the participants wore audio headsets equipped
with microphones to achieve real-time verbal intercommunication, as shown in Figure 6.

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - July 2018