Skip to main content

Table 9 Average results and standard deviation for the Rule-based (RB) and Neural Network (NN) methods for each exercise (E1, E2, and E3) with LOSO and LOES cross-validation

From: A low-cost virtual coach for 2D video-based compensation assessment of upper extremity rehabilitation exercises

  Precision Recall \(F_1 score\) Hamming loss
Rule-based (RB) Approach
 Leave-One-Subject-Out (LOSO) cross-validation
  \(\text {E1}\) 0.75 ± 0.14 0.78 ± 0.12 0.76 ± 0.12 0.11 ± 0.07
  \(\text {E2}\) \(0.54 \pm 0.17\) \(0.65 \pm 0.17\) \(0.59 \pm 0.16\) \(0.20 \pm 0.08\)
  \(\text {E3}\) \(0.69 \pm 0.27\) \(0.71 \pm 0.26\) \(0.70 \pm 0.26\) 0.13 ± 0.11
Neural Network (NN) based Approach
 Leave-One-Subject-Out (LOSO) cross-validation
  \(\text {E1}\) \(0.71 \pm 0.23\) \(0.70 \pm 0.25\) \(0.70 \pm 0.24\) \(0.18 \pm 0.15\)
  \(\text {E2}\) 0.73 ± 0.21 0.73 ± 0.19 0.73 ± 0.19 0.15 ± 0.11
  \(\text {E3}\) 0.80 ± 0.22 0.80 ± 0.21 0.80 ± 0.22 \(0.14 \pm 0.14\)
 Leave-One-Exercise-Out (LOEO) cross-validation with E1 and E2
  \(0.78 \pm 0.05\) \(0.81 \pm 0.01\) \(0.80 \pm 0.02\) \(0.12 \pm 0.01\)
  1. The results in bold correspond to the best classifiers' performance for the different metrics for each exercise
  2. F1 score is a measure of accuracy