Identification of Acute Myocardial Infarction from Left Ventricular Wall Rupture Using ResNet 18-Deep Active Learning Algorithms

Acute myocardial infarction (AMI) is a heart muscle ischemia caused by blockage or narrowing of coronary arteries, leading to left ventricular wall rupture (LVWR).Diagnosing AMI is challenging due to improper border and edge enhancement, segmentation, and classification.To address the above-mentioned, a deep denoised convolutional neural network (DnCNN) is applied to enhance the edge and boundary regions of the myocardium.A ResNet 18-based deep active curriculum learning (DACL) model is proposed to classify MI or non-MI patients by left ventricular wall rupture.The model is trained with a few samples to detect MI and dynamically updates the number of samples in the training dataset.The adaptive sampling strategy efficiently classifies myocardial infarction in the HMC-QU dataset, achieving a sensitivity of 98.2%, a specificity of 97.3%, an accuracy of 98.5%, and an AUC of 99.6%.