BACKGROUND
In most previous studies, the acceleration sensor is attached to a fixed position for gait analysis. However, if it is aimed at daily use, wearing it in a fixed position may cause discomfort. In addition, since an acceleration sensor can be built into the smartphones that people always carry, it is more efficient to use such a sensor rather than wear a separate acceleration sensor.
OBJECTIVE
We aim to distinguish between hemiplegic and normal walking by using the inertial signal measured by means of an acceleration sensor and a gyroscope.
METHODS
We used a machine-learning model based on a convolutional neural network to classify hemiplegic gaits, and used the acceleration and angular velocity signals obtained from the system freely located in the pocket as inputs without any pre-processing. We evaluated the performance of the developed model by means of a clinical trial including a walking test on 42 subjects (57.8 ± 13.8 years, 165.1 ± 9.3 cm, 66.3 ± 12.3 kg) including 21 hemiplegic patients.
RESULTS
The developed model showed an accuracy of 0.86, a precision of 0.90, a recall of 0.99, an area under the receiver operating characteristic curve of 0.91, and an area under the precision-recall curve of 0.97.
CONCLUSIONS
We confirmed the possibility of distinguishing hemiplegic gaits by means of a machine-learning model without additional pre-processing or feature extraction, using a 6-axis inertial signal measured at random locations in a pocket, like a smartphone.
CLINICALTRIAL
SCH2016-130