Table 2.

Comparing results of Mocap-based style/subject recognition (HDM05 dataset). The average accuracy is reported concerning the cited works vs. our approach—EHECCO+SVM.

Method	Accuracy (%)
SPDNet [40]	61.45
SE [41]	70.26
SO [42]	71.31
LieNet [31]	75.78
Seq2Im+SVM [27]	70.70
Seq2Im+KNN [27]	66.82
Seq2IM+RF [27]	80.62
Seq2Im+CNN (fine-tuning) [27]	83.33
EHECCO+SVM	88.80