Prediction of Activity Energy Expenditure Using Kinect™

• Main Authors: Riotaro Sananta, 鍾立善
• Other Authors: Chiu-Hsiang Lin
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/14875813504925489551

碩士 === 國立臺灣科技大學 === 工業管理系 === 101 === A physically active lifestyle will enhance the feelings of energy, well-being, quality of life, and cognitive function and is associated with lower risk of cognitive decline and dementia. Good personal health and fitness in general should be monitored by one’s energy expenditure (EE). The development of active video game (AVG) has the potential to encourage people to spend more active leisure time. By utilizing the biomechanical principle, Kinect™ body joint detection could also facilitate the calculation of human EE over time. There are two possible equations that could be used to estimate EE: mechanical energy (KineticE) and work (WorkE). Participants performed four task sessions, playing Wii™ Mario Kart, playing Wii™ Sports Golf, playing Wii™ Sports Boxing, and playing Wii™ Dance Dance Revolution Hottest Party 2. Validation result on holdout samples shows significant correlation between AEEkinetic (R-squared=0.254) and AEEwork (R-squared=0.643) with indirect calorimetry. Overall, AEEwork model can be utilized as a new alternative for predicting activity energy expenditure.