Effects of Image Size on Deep Learning

In this work, the best size for late gadolinium enhancement (LGE) magnetic resonance imaging (MRI) images in training dataset was determined to optimize deep learning outcomes. Non-extra pixel and extra interpolation algorithms were used determine new of LGE-MRI images. A novel strategy introduced handle masks remove class labels interpolated ground truth (GT) segmentation masks. The expectation maximization, weighted intensity, a priori information (EWA) algorithm quantification myocardial infarction (MI) automatically segmented Arbitrary threshold, comparison sums, sums differences are methods estimate relationship between semi-automatic or manual fully automated results. MI results found be closer case bigger LGE MRI (55.5% results) than smaller (22.2% results).