Comparing of CNN Algorithms for Feature Extraction and Classification on Cardiac MRI

Deep Learning (DL) has been incorporated into image processing, including cardiac MRI images to address Machine Learning (ML) issues, the best model based on research evaluation is still up for debate among experts in the field. The diversity of datasets has characteristics that are not always in accordance with the objectives of the research to be carried out. Each of the existing models each has a different performance achievement. This review contributes to presenting various datasets and DL models related to their application to cardiac MRI images. The proposed classification method goes through several main stages, including object detection to obtain the region of interest (RoI), segmentation to obtain feature maps, classification with a fully connected layer (FCN).