Each line of the csv file is a 99-dimensional vector representing kinematics and positions of markers. It's followed by 2 0-1 numbers, representing a binary flag of events: foot strike and foot off in the 'target leg', i.e. the leg for which we are looking for gait events. All angles are represented by 3 numbers (yaw, pitch, roll). Positions of markers are represented as 3-dimensional points X, Y, Z, relative to the pelvis (except for the pelvis itself). First 15 numbers represent angles of: hip, knee, ankle, pelvis, and foot progression in the target leg. Next 15 numbers represent angles in the opposite leg. Next 15 numbers represent positions of the following markers: ankle, toe, knee, anterior superior iliac spine, heel in the target leg. Then following next 15 numbers represent the opposite leg. Next 30 numbers are discreet derivatives of markers in the target and opposite legs. Finally the next 9 numbers represent postions, velocities and accelerations of the pelvis. Please refer to https://github.com/kidzik/event-detector-train/blob/master/read.py#L17-L130 for details on how exactly were these vectors constructed.  View License