This review research addresses the critical challenge of automatically segmenting cardiac components in medical imaging. We provide a comprehensive review of the cutting-edge deep learning (DL) architectures for this task, highlighting their innovative concepts and effectiveness. The review includes a wide range of datasets used for training and assessment of these designs, describing their properties and the particular modalities (e.g., MRI, CT, echocardiography, and CCTA) they represent. Our goal is to provide a thorough review of the most recent methods for segmenting the heart’s anatomy by methodically examining these elements. We also address the shortcomings and difficulties encountered by existing approaches, such as dataset variability, applicability to other imaging modalities, and computing complexity. This study adds to the field by summarizing current developments and pointing out areas that need more investigation. As a result, it offers practitioners and researchers working on cardiac imaging and analysis useful information. The survey’s goal is to discuss the challenge of autonomously segmenting heart architecture. Analyzing the deep learning architectures, modalities, and datasets used. Examining and debating the features of existing DL datasets and systems. A summary of innovative techniques and their effectiveness.