Faculty Mentor(s)

William Seffens, Paula Seffens, Sam Fouche, Molly Martin

Campus

Gainesville

Proposal Type

Poster

Subject Area

English/Communications

Location

Nesbitt 3110

Start Date

23-3-2018 11:00 AM

End Date

23-3-2018 12:00 PM

Description/Abstract

PURPOSE: Many innovative information technology applications use gestures as input that span a variety of platforms, from touch screen of a smart phone to natural input. Visual Gesture Builder (VGB) for Kinect, a data-driven machine learning solution for gesture detection, was used to capture poses with high accuracy. This gesture analysis technology is explored for incorporation into exergames for personalized medical interventions using yoga as therapy (YT). Research goal was to test whether machine learning algorithm in basic computer video exergame could assess skill acquisition in targeted populations to promote healthy physical activity. METHODS: Convenience sample of 20 adult students, male and female, were briefly instructed and shown poses to perform, while recorded by a Kinect attached to a PC. The capture resulted in raw files 10-20 GB in size. RESULTS: For comparison, we recorded 6 instructors in a series of postures using Kinect Studio. Recorded clips were tagged in all of the frames that defined a gesture by consensus of two researchers. Default settings in Kinect VGB produced solutions with high True Positives (99.5%) and low False Positives (0.03%) for most postures sampled. We measured posture alignment over the course of a 10-week period in an IRB approved study. Depth stream and skeleton coordinates for the 20 participants were acquired and analyzed against the previous trained solution. Analysis of summary statistics was done for the mountain pose comparing initial, mid-session, and final session captures with Sensitivity and Informedness showing the most significant t-test between initial and final. Sensitivity went from 0.79 to 0.90, while the expert test clip scored 0.94. CONCLUSION: Gesture analysis for alignment training may be a useful tool for the development of home and clinical YT for hard to reach populations. The experimental exergame developed here provides a tool that scores performance of postures and provides improvement metrics. Prior research by others has shown that even short-term yoga based lifestyle interventions were efficacious in weight loss, inflammation and stress and positively influenced cardiovascular risk factors. Our plans are to target special populations with YT, and study the potential effects of body mass and age on posture alignment and limb stretch.

Share

COinS
 
Mar 23rd, 11:00 AM Mar 23rd, 12:00 PM

41. Gesture Analysis of Yoga Poses for Exergame Using Machine Intelligence

Nesbitt 3110

PURPOSE: Many innovative information technology applications use gestures as input that span a variety of platforms, from touch screen of a smart phone to natural input. Visual Gesture Builder (VGB) for Kinect, a data-driven machine learning solution for gesture detection, was used to capture poses with high accuracy. This gesture analysis technology is explored for incorporation into exergames for personalized medical interventions using yoga as therapy (YT). Research goal was to test whether machine learning algorithm in basic computer video exergame could assess skill acquisition in targeted populations to promote healthy physical activity. METHODS: Convenience sample of 20 adult students, male and female, were briefly instructed and shown poses to perform, while recorded by a Kinect attached to a PC. The capture resulted in raw files 10-20 GB in size. RESULTS: For comparison, we recorded 6 instructors in a series of postures using Kinect Studio. Recorded clips were tagged in all of the frames that defined a gesture by consensus of two researchers. Default settings in Kinect VGB produced solutions with high True Positives (99.5%) and low False Positives (0.03%) for most postures sampled. We measured posture alignment over the course of a 10-week period in an IRB approved study. Depth stream and skeleton coordinates for the 20 participants were acquired and analyzed against the previous trained solution. Analysis of summary statistics was done for the mountain pose comparing initial, mid-session, and final session captures with Sensitivity and Informedness showing the most significant t-test between initial and final. Sensitivity went from 0.79 to 0.90, while the expert test clip scored 0.94. CONCLUSION: Gesture analysis for alignment training may be a useful tool for the development of home and clinical YT for hard to reach populations. The experimental exergame developed here provides a tool that scores performance of postures and provides improvement metrics. Prior research by others has shown that even short-term yoga based lifestyle interventions were efficacious in weight loss, inflammation and stress and positively influenced cardiovascular risk factors. Our plans are to target special populations with YT, and study the potential effects of body mass and age on posture alignment and limb stretch.