TY - GEN
T1 - Automatic Step Detection of Tandem Gait Test in Patients with Vestibular Hypofunction Using Wearable Sensors
AU - Huang, Yi Ju
AU - Liu, Chien Pin
AU - Ting, Kuan Chung
AU - Hsieh, Chia Yeh
AU - Liu, Kai Chun
AU - Chan, Chia Tai
N1 - Publisher Copyright:
© 2022 Asia-Pacific of Signal and Information Processing Association (APSIPA).
PY - 2022
Y1 - 2022
N2 - The tandem gait test is a common examination in testing individuals with possible neurologic diseases, such as motor neurons or cortex problems, and vestibular disorders. A reliable step detection method is essential to capture these step-based features to assess vestibular hypofunction. However, most studies still relied on manual labeling or a simple threshold-based peak detection approach to obtain step information. These methods often suffer issues in time consuming and reliability. This study proposes an automatic and accurate step detection algorithm for the tandem gait test using a dynamic threshold. This technique could filter noises and adapt to individual step patterns. The proposed algorithm was validated on a dataset including 15 healthy subjects and 62 patients with peripheral vestibular disorders. The results show that using the developed step detection approach with the shank-worn sensor has the best performance, which achieves 99.8%, 100%, 99.8% and 99.9% in accuracy, recall, precision and F1-score, respectively.
AB - The tandem gait test is a common examination in testing individuals with possible neurologic diseases, such as motor neurons or cortex problems, and vestibular disorders. A reliable step detection method is essential to capture these step-based features to assess vestibular hypofunction. However, most studies still relied on manual labeling or a simple threshold-based peak detection approach to obtain step information. These methods often suffer issues in time consuming and reliability. This study proposes an automatic and accurate step detection algorithm for the tandem gait test using a dynamic threshold. This technique could filter noises and adapt to individual step patterns. The proposed algorithm was validated on a dataset including 15 healthy subjects and 62 patients with peripheral vestibular disorders. The results show that using the developed step detection approach with the shank-worn sensor has the best performance, which achieves 99.8%, 100%, 99.8% and 99.9% in accuracy, recall, precision and F1-score, respectively.
UR - http://www.scopus.com/inward/record.url?scp=85146280356&partnerID=8YFLogxK
U2 - 10.23919/APSIPAASC55919.2022.9979880
DO - 10.23919/APSIPAASC55919.2022.9979880
M3 - Conference contribution
AN - SCOPUS:85146280356
T3 - Proceedings of 2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2022
SP - 1550
EP - 1555
BT - Proceedings of 2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2022
Y2 - 7 November 2022 through 10 November 2022
ER -