Fig. 4From: Development and clinical validation of inertial sensor-based gait-clustering methods in Parkinson’s diseaseSemi-automatic stride segmentation of a sequence of a 4x10 m gait test signal. (Top) Strides automatically segmented by the DTW algorithm using only the GZ axis as a template (blue boxes). The minima in the GZ signal (red crosses) were used to define stride borders for missing or wrongly segmented strides. (Bottom) Fully segmented sequence of a 4x10 m gait test signal. Missing or wrongly segmented strides were manually added or corrected based on the minima in the GZ signalBack to article page