Table 9.

Comparison of the classification result from Schuldhaus et al. [13] and our own evaluation approaches segmented_lab, window_lab and window_field for the SVM classifier. Schuldhaus et al. [13] classified full-instep versus side-foot kicks (marked with *) on segmented laboratory data, whereas our approaches classify shots and passes in general. In segmented_lab, segmented events extracted from lab data are used for evaluation. window_lab and window_field apply moving windows, peak detection for candidate selection and classification to laboratory and field data, respectively. In accordance with Schuldhaus et al. [13] we reported the class specific sensitivity scores.

Class Label	Schuldhaus et al. [13]	Segmented_Low	Window_Low	Window_High
Null	-	93.3%	91.7%	83.2%
Pass (Side-Foot *)	97.9%	21.7%	23.5%	20.9%
Shot (Full-Instep *)	95.6%	38.8%	37.9%	36.0%