VERG -> Virtual Patients ->
Overview
|
Overview |
System Description The
experiences occur in an examination room.
The system is composed of two computers, four cameras for tracking the
student’s posture, a data projector, a wireless microphone, and Dragon
Naturally Speaking 9 for speech recognition.
The total cost of the hardware involved is less than $8000, and the use of commodity components
makes wide-spread adoption a realistic goal. The user’s
posture is approximated by tracking IR retro-reflective tape on a ball cap,
chair, and finger. Students can use
the hand tracking to localize the VP’s pain with simple pointing
gestures. A wireless microphone
captures audio input (speech recognition performance is about 70% matching
for utterances). |
|
|
Experience The VHs’ gestures
and audio responses were drafted by teaching medical faculty. Users knock on the exam room door, enter,
and see VHs projected life-size on the exam room wall. The user converses with the VP for
ten-minutes. Several scenarios have
been created including focusing on acute abdominal pain, breast mass, and
blurred vision. Evaluation Since August 2004, studies have been conducted to
develop, evaluate, and validate the system.
Medical, nursing, and physician assistant students (n > 150) participated. The system is being installed and tested at
the Medical College of Georgia, Videos of interactions are available upon request. |
What
we think we know about virtual humans: ·
Interaction with a virtual patient is
validated. Expert observer ratings of virtual patient
interactions are correlated (r=0.49)
with standardized patient interactions. ·
Conversation content with a virtual patient is
similar to that with a standardized patient, even if the method be might more robotic. ·
Although the interaction with a VP is not identical
to a SP, some educational objectives can be achieved. High
level concepts, e.g. empathy, require further research. ·
The patient-doctor interaction is a usable platform
to study virtual humans, despite current technology compromises. ·
Natural interaction with virtual humans is
important for teaching communication skills. |
|
|
Ongoing
Research Directions Research and user studies are currently underway
to: ·
Validate the
use of virtual humans to teach communications and interpersonal skills ·
Characterize a
virtual human interaction ·
Expose
students to abnormal findings, e.g. neurological, psychomotor, and emotional ·
Elicit real
world biases (e.g. racial/ethnic, gender, age, intelligent, weight, and
accent) ·
Analyze and
visualize the interaction with a virtual human to enable student
self-reflection ·
Visualize,
categorize, and evaluate the signals from a virtual human-human interaction ·
Use highly
interactive virtual environments that allow real tools to perform virtual
exams. ·
Evaluating the
effect of mixed reality interaction on presence and co-presence. ·
Measure the
impact of system performance on perception, presence, and co-presence. |