Schematic overview of the procedure of pattern recognition (1 of the 36 signals is shown).

<p>A) Systematic variation in template window length using linear interpolation, example given of 3 of the in total 17 variations in window length, B) Pattern recognition based on 2D convolution, performed for every template window separately, C) Cross-correlation coefficient functions, calculated during pattern recognition, example given of 3 of the in total 17 cross-correlation coefficient functions. Local maxima were identified per cross-correlation coefficient function (marked with red circle) and local maxima with highest cross-correlation of all cross-correlation functions were identified (marked with red square), D) Overview of epochs where the template is recognized, including start and endpoints (in data points, DP), cross-correlation coefficients and relative window length. Red lines mark the places in the sequence of recordings of multiple standardized daily activities were the template is recognized. DP = data point.