File
E. Park1, S. Cheon2, J. Kim3, J.Y. Hwang4, H. Nam5
1University of North Georgia, Department of Physical Therapy, Atlanta, United States, 2Youngsan University, Department of Physical Therapy, Yangsan, Korea (Republic of), 3Kent State University, Department of Computer Science, Kent, United States, 4Kent State University, School of Fashion, Kent, United States, 5CHA University Medical Center, Department of Rehabilitation of Medicine, Gumi, Korea (Republic of)
Background: Gait is the most important daily activity which is highly related to human health. Monitoring gait in daily life to detect abnormal events can help to reduce the negative impact on mechanical abnormality. The pes planus (a.k.a. flatfoot) affect the alignment on foot, ankle, leg, pelvis, and spine during gait.
Purpose: Wearable devices to detect changes of alignment with flatfoot during gait were focused on the foot movement with simple sensing devices, such as force sensors and inertial measurement unit (IMU) sensors. However, those devices are not accurate to address ankle movement in the gait cycle. In this study, we developed an integrated smart wearable gait monitoring device with three sensors at the front foot, the rear foot, and a flexible sensor at the ankle. We also have explored the more accurate and dynamic flatfoot detecting method based on a dynamic sensing window and a deep neural network with scaled principal component analysis (PCA).
Methods: We have tested 24 subjects (12 flatfoot subjects). These diagnoses were conducted by an expert clinician based on the navicular drop test that compares the difference of the length from navicular to ground between weight-bearing and non-weight-bearing. The collected data were processed based on PCA and DNN for classifying the flatfoot and non-flatfoot subjects.
Results: This study shows that the proposed sensing devices were worn comfortably and the proposed DNN model outperformed the other five common classifier algorithms considered and the area under the curve (AUC) value of our method was 87.1%.
Conclusion(s): Our sensing device can provide a comfortable sensing environment and accurate performance to investigate the progress of flatfoot. Our methodology also can be a prescreening tool for possible gait problems in daily life.
Implications: This proposed device can be used not only for patients who are training to improve gait problems with flatfoot but also applied to detect the progress of flatfoot.
Funding, acknowledgements: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government. (No. NRF-2017R1C1B5076194).
Keywords: Wearable Sensor, Gait Monitoring, Flatfoot
Topic: Innovative technology: information management, big data and artificial intelligence
Did this work require ethics approval? Yes
Institution: Youngsan University
Committee: IRB Committee
Ethics number: YSUIRB-2017-HR-028-02
All authors, affiliations and abstracts have been published as submitted.
