Computerized image analysis workflows have facilitated the clinicians/researchers with improved characterization of biological structures for comprehending obscure abnormalities. Previously, the workflows employing classical image analysis methods were explicitly designed for problem-specific solutions. Since the inception of deep learning as a powerful recognition method, the research interest has shifted from problem-specific solutions to increasingly problem-agnostic solutions that rely on learning from data. In particular, convolutional neural networks (CNNs) have rapidly become a primary choice due to its promising results. Given the prevalence of deep learning methods, I will briefly discuss the underlying concepts of CNN and also highlight their applications in multiple tasks of an automated workflow, e.g., pre-processing, segmentation, classification, etc.