Template Matching - Discriminator
The template matching discriminator script is designed to match a data series against two templates, and also determine which is a stronger match, thereby associating a section of input data with only one template (the stronger match), even when there is a match for both templates in the same region. This allows the user to distinguish between two similar actions in a data set, i.e. running vs. walking. For example, accelerometer data from the foot. The shape of one step while walking is very similar to the shape of one step while running. Therefore, during a simple Template Matching calculation, matches for either walking or running templates may appear across the entire data series. The Template Matching Discriminator examines the correlation data for areas of apparent template match overlaps, and performs a comparison of the correlations to determine which template is better matched to the area. Only one template will return as a true match for that area. Thus the output of this calculation is a Matches Output for Template A and a Matches Output for Template B, with no overlaps between them.
Parameters