Article Preview
TopBrief Description Of The Technology
Prior to describing the results, we begin with a description of the technology used in our virtual patient work. We define virtual patients as embodied conversational agents – computer-generated virtual characters that are controlled by a computer algorithm with whom users can conduct a conversation. This conversation is modeled as a knowledge-base of possible answers the virtual patient can give to input from the user. Input from the user can be in the form of questions or gestures.
Users may type, speak or choose from a list their questions to the virtual patient. These questions are processed by the system using a keyword-based algorithm. This algorithm chooses the best answer from the knowledge-base of the virtual patient. The virtual patient then speaks the answer using prerecorded audio.
Users may also interact with the virtual patient using gestures to perform physical examinations on the virtual patient. These user gestures can be recognized by the system using: infrared cameras tracking retro-reflective markers, a Nintendo Wiimote, a Microsoft Kinect, metaphors using keyboard and mouse, or customized sensors that resemble body parts of the virtual patient.
Depending on the requirements of each training scenario, virtual patients can be displayed as life-size on a wall using a projector or on a TV, or in smaller sizes using desktop monitors or laptop displays (see Figure 1).
Figure 1. Images of ways in which to interact with virtual patients
TopReview Of Existing Findings
Historical perspectives on technological advancements are not always useful as newer technologies often replace prior conditions, outcomes, problems, and solutions. However, this review documents important findings that drive current and future planning; they set the stage for others interested in a longitudinal perspective not just on the project, but also how to build on our successes, shortcomings, and learnings.