Fig. 3

Overview of the explanation framework in XplainScar. We feed the entire JH dataset as the training set to Kernel Explainer and use the Shapley values for each ECG from the UCSF dataset for LV scar prediction.
Overview of the explanation framework in XplainScar. We feed the entire JH dataset as the training set to Kernel Explainer and use the Shapley values for each ECG from the UCSF dataset for LV scar prediction.