Drinking Gesture Spotting and Identification Using Wrist-Worn Inertial Sensor

• Main Authors: Liu-Hsuan Chen, 陳劉軒
• Other Authors: Chia-Tai Chan
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/44pfs5

碩士 === 國立陽明大學 === 生物醫學工程學系 === 105 === Stroke is the main serious long-term disability in the world. There are 77% of stroke patients suffering the upper extremity disability and loss of function of upper limb motor. Furthermore, the disability and function loss limit independence and social participation of patients. Rehabilitation exercise is one of most medical treatment for the patients to recover the lost function of upper limb. In order to understand and realize the progression of the rehabilitation, many clinical approaches are generally utilized to assess upper-limb performance include self-reports, questionnaires, and evaluation tools, i.e. Modified Rankin Scale. However, these typical assessment approaches suffer some issues, such as pervasiveness, time consuming, human resource limitation. With the progress of the Micro-Electromechanical Systems (MEMS), various studies have shown that the upper limb monitoring system using wearable computing and technologies has the potential to objective, long-term and unobstructive monitor and assess performance of upper-limb movement during daily living. It also can support clinical profession to keep track of the progress and provide the adequate assistance for the patients. In this work, we develop drinking gesture monitoring system using wrist-worn inertial sensor for assessing performance of upper-limb movement during daily living. In the proposed drinking gesture monitoring system, the Support Vector Machine-based drinking gesture spotting model is proposed to observe the drinking gesture during daily living. The rule-based transition detection model is proposed for identification of elementary motions including extension and flexion.