This study investigates automating the analysis of myocardial native T1 mapping images from cardiovascular magnetic resonance (CMR) using a fully convolutional neural network (FCN). This process currently relies on time consuming manual segmentation. The FCN was trained on 210 manually segmented datasets and validated on 455 additional patient scans. It rapidly segmented myocardium with a high Dice coefficient and produced global and regional T1 values that closely matched expert measurements, with strong correlations (r = 0.72-0.82). The authors iterated that this approach reduces observer variability and the tedious nature of manual analysis, although there are still challenges with poor image quality and greater slice level variability.