Validation of automatic detection of gait phases

This dataset contains a back axis video screen recording of DIERS formetric III 4D™, DICAM v3.7Beta analyzing system and the exported corresponding raw data in an Excel sheet.
Encoding of gait phases was as follows: We started to measure a gait cycle (encoded with 1 in the exported output) with the first full Initial Contact of the right foot which is also the start of the encoding for the stance phase right (1). Hence, the start of the next gait cycle is the next IC of the right leg (2) and so forth (3 …). At the first right IC the left leg is supposed to be in Terminal Stance, therefore still on the ground but without having had its first IC. This and only this context is encoded by 0, the full stance period after IC left is analogously encoded by 1. Both legs swing phases are marked with - in the data export. Along their assigned time stamp the respective gait phases (stance and swing phases left/right; full gait cycles) were encoded timely synchronized into the raw data of the spinal model.
Validation of automatic detection of gait phases was done by viewing video recordings of the back and lateral axis cameras in DICAM in which single frames of the video recordings (also containing the same time stamp / frame as the spinal model data output) were checked for face valid results compared to the automated detection and its encoding into the spinal model data. A data sample Excel sheet of one single measure at 5 km/h including one spinal parameter (Rotation Pelvis) can be used as a reference for this process. In all 3 gait cycles the visually identifiable moment in time of Initial Contact was in the range of ± 3 Frames / 3/60 Hz compared to frame number in the exported raw data.