In cardiac MRI, left ventricle (LV) segmentation typically follows the identification of a short-axis slice range, which may require a manual procedure. The standard cardiac image processing guidelines indicate the importance of correct identification of the slice range. In this study, we investigate the feasibility of deep learning in automatically classifying the slice range. Images were classified into one of three categories: out-of-apical (OAP), apical-to-basal (IN), and out-of-basal (OBS). We developed our in-house user interface to label image slices into one of the three categories for learning. We evaluated the performance of the models, fine-tuned from seven popular deep CNNs.