Figure 9.

Gait reference system based on vision. (a) The process of extracting the joint angles of a leg: I. Distribution of leg markers. II. Skeleton model of the right leg. III. Extracted joint angles. (b) The basis of gait classification labels in machine learning [Perry’s method of dividing the period of gait, LR (loading response), MS1 (mid-stance), TS1 (terminal stance), PS (pre-swing), IS (initial swing), MS2 (mid-swing), TS2 (terminal swing). (c) Angle data of three joints (hip, knee, and ankle) extracted from the visual system after data processing. (d) The theoretical classification and the machine learning forecast classification of the phase of visual leg angles during a walking gait cycle.