The paper provides an overview of deep learning in medical imaging, highlighting its rapid growth and significance. It begins by discussing the increasing use of machine learning (ML) in medical imaging, particularly in computer-aided diagnosis (CAD), radiomics, and medical image analysis. The introduction of deep learning, a subset of ML, in the computer vision field, starting with the 2012 victory of a deep-learning approach in the ImageNet Classification competition, is emphasized. This event marked the beginning of deep learning's explosive growth across various fields. The paper outlines the key changes in ML before and after the introduction of deep learning, focusing on the shift from feature-based ML to image-based ML. Deep learning's power lies in its ability to learn directly from image data without the need for feature extraction or object segmentation. Two major models in deep learning are discussed: massive-training artificial neural networks (MTANNs) and convolutional neural networks (CNNs). These models have both similarities and differences, with MTANNs generally being more efficient and requiring fewer training cases compared to CNNs. The applications of these models in medical imaging are explored, including their use in reducing false positives in CAD, distinguishing between benign and malignant lesions, and improving the accuracy of image analysis. The paper concludes by emphasizing the promising future of deep learning in medical imaging, predicting that it will become the mainstream area in the field over the next few decades.The paper provides an overview of deep learning in medical imaging, highlighting its rapid growth and significance. It begins by discussing the increasing use of machine learning (ML) in medical imaging, particularly in computer-aided diagnosis (CAD), radiomics, and medical image analysis. The introduction of deep learning, a subset of ML, in the computer vision field, starting with the 2012 victory of a deep-learning approach in the ImageNet Classification competition, is emphasized. This event marked the beginning of deep learning's explosive growth across various fields. The paper outlines the key changes in ML before and after the introduction of deep learning, focusing on the shift from feature-based ML to image-based ML. Deep learning's power lies in its ability to learn directly from image data without the need for feature extraction or object segmentation. Two major models in deep learning are discussed: massive-training artificial neural networks (MTANNs) and convolutional neural networks (CNNs). These models have both similarities and differences, with MTANNs generally being more efficient and requiring fewer training cases compared to CNNs. The applications of these models in medical imaging are explored, including their use in reducing false positives in CAD, distinguishing between benign and malignant lesions, and improving the accuracy of image analysis. The paper concludes by emphasizing the promising future of deep learning in medical imaging, predicting that it will become the mainstream area in the field over the next few decades.