A Real-Time System for Body Visualization During Whole-Body Motion, 07-R9675

Printer Friendly Version

Principal Investigators
Joe D. Wilson
Warren C. Couvillion
Ryan C. Logan

Inclusive Dates:  01/01/07 – 01/01/09

Background - Humans are generally poor at visualizing their bodies using their kinesthetic sense alone, especially when in action, making it difficult to learn or practice motor skills. Research has shown that visual cues can improve motor skill development, and a variety of techniques have been applied to whole-body visualization including the use of mirrors, video displays, motion capture and video capture/analysis. However, none of these techniques provides real-time feedback while the user performs a motion in a natural manner. In comparison, training methods that make use of post-performance assessment, such as video analysis, are problematic because short-term kinesthetic memory is very brief.

Approach - The objective of this project was to develop a prototype system. The system comprises a low-cost, see-through head-mounted display (HMD), cost-effective web camera, laptop and specially designed software to allow users to view captured video of the body augmented with visual cues in real time as they perform natural motions. Software required to capture video and overlay visual cues was developed using the SwRI Graphics Interface Library (GraIL™) software developed on a previous internal research effort.

Accomplishments - A system composed of a laptop computer, web video camera and a lightweight HMD was developed and utilized to allow users to visualize their bodies from an external perspective augmented with visual cues. The system processes video from a camera and provides an interface for presenting and augmenting the resulting HMD view. The system provides static visual indicators, which overlay the video. Lines, circles, arcs, and a grid can be placed on the video window and adjusted for scale and position to provide cues regarding proper body posture/positioning. Using this approach, a proof-of-concept application was developed allowing participants to work with three predefined shoulder exercise routines. The participant's arm position and average velocity during exercise repetitions are automatically collected and saved to a database. The system was set up in a rehabilitation facility where data was collected on shoulder surgery patients to evaluate the basic approach.

These images show the participant's view (left) as seen through the HMD, and the operator's view (right) including controls and participant display.

2009 Program Home