Myocardial disease refers to diseases that affect the heart muscle and can lead to heart failure if not diagnosed and treated early. In fact, early diagnosis by segmental MRI of the heart is important to improve patient outcomes. In this study, a deep learning-based approach is proposed which uses DeepLabv3 for semantic segmentation and ResNet for feature extraction to identify myocardial abnormality from MRI scans. While the porous spatial pyramid pool of DeepLabv3 can capture many features, ResNet can extract deep and hierarchical features that are important to distinguish tissue from healthy myocardium. The system is designed to improve the segmentation accuracy of important cardiac structures including myocardium and provide a reliable estimation of myocardium at an early stage. Initial results show that the model outperforms traditional methods in terms of segmentation accuracy and has the potential to improve the clinical decision-making and personalized treatment strategies.