Deep learning myocardial infarction segmentation framework from cardiac magnetic resonance images

Mugahed A. Al-antari, Zakarya Farea Shaaf, M. Mahadi Abdul Jamil, Nagwan M. Abdel Samee, Reem Alkanhel, Muhammed Talo, Zaid Al-Huda

Abstract: Highlight•A proposed AI-based myocardial infarction (MI) segmentation framework is utilized using short-axis LGE-MR images.•The adopted framework with ReU-Net demonstrates outstanding segmentation performance with an overall accuracy of 88.48% and 35 Frames per Second.•A thorough investigation and ablation study are conducted to optimize the best AI segmentation model for MI and MVO segmentation tissues among LV and Myo normal ones.