Figure 3
From: Tool-tissue force segmentation and pattern recognition for evaluating neurosurgical performance
Performance comparison between LSTM and FTFIT models with hand-crafted feature combinations in surgical skill prediction and task recognition. Different combinations of hand-crafted features, i.e., no feature, selected set of features identified through KNN and XGBoost feature importance ranking, and a subset of features (Duration Force, Range Force, Entropy, and Heterogeneity; which were consistent with the features presented in SmartForceps performance dashboard1) have been compared.
