Table 2.

Model performance metrics (%) for the diagnosis of acute myocarditis in CMR.

Models	Accuracy	Precision	Recall	Specificity	F1-score	AUC
CNN-KCL [Sharifrazi et al. (45)]	97.41	97.6	95.7	98.56	96.5	97.05
RLMD-PA [Moravejj et al. (46)]	88.6	84	86.3	90.1	85.1	N/A
TNT [Jafari et al. (47)]	99.68	99.47	99.59	99.72	99.53	99.94
Inception v4 [Jafari et al. (47)]	99.7	99.44	99.69	99.71	99.57	99.94

NB for RLMD-PA, G means (88.2%) instead of AUC was calculated.

CNN-KCL, convolutional neural network k-means clustering; RLMD-PA, reinforcement learning-based myocarditis population-based algorithm; TNT, turbulence neural transformer.