Evaluation of Optical Myography Sensor as Predictor of Hand Postures

Autor(es)

Wu, Yu Tzu ^[1] ; Fujiwara, Eric ^[1] ; Suzuki, Carlos K. ^[1]

Afiliação do(s) autor(es)

^[1] Univ Estadual Campinas, Sch Mech Engn, Lab Photon Mat & Devices, BR-13083860 Campinas, SP - Brazil

• Tipo de documento: Artigo Científico
• Fonte: IEEE SENSORS JOURNAL; v. 19, n. 13, p. 5299-5306, JUL 1 2019.
• Citações Web of Science: 1

Resumo

Optical myography stands as one of many techniques to assess hand postures. As a combination of computer vision and muscular activity analysis, it differs from conventional gesture recognition techniques, such as instrumented gloves or optical tracking, by not relying on the existence of a healthy hand, so it is able to detect the hand motion as well as the motion intent. In this aspect, optical myography is like well-established myographic approaches, such as surface electromyography or force myography, but it is simpler, more comfortable, and inexpensive. This recent technology is hereby evaluated upon the construction of a feasible and low-cost sensor that monitors both the front and the back of the forearm. The results are organized into two sections: the first validates the sensor and the second evaluates its performance as a predictor of eight static postures, including the thumb and the fingers motion. In the end, the sensor proved to be comparable to more mature techniques with an F-score of similar to 92.2% and similar to 71.5% for front- and back-side analysis, respectively. (AU)

• Processo FAPESP: 17/25666-2 - Desenvolvimento de sensor de fibra óptica para medição de sinais de miografia de força aplicado a interfaces humano-robô
• Beneficiário: Eric Fujiwara
• Modalidade de apoio: Auxílio à Pesquisa - Regular