This article provides an overview of deep learning in medical imaging, highlighting its history, development, and applications. Deep learning, a subset of machine learning (ML), has gained significant attention due to advancements in big data, enhanced computing power, and novel algorithms. The introduction of artificial neural networks (ANN) in the 1950s was limited by issues such as vanishing gradients, overfitting, and insufficient data. However, recent breakthroughs have addressed these limitations, leading to impressive performances in various fields, including medical imaging. The article discusses the evolution of ML from traditional methods to deep learning, emphasizing the importance of feature extraction and the challenges of training deep architectures. Convolutional neural networks (CNNs) and recurrent neural networks (RNNs) are highlighted for their applications in image segmentation, registration, automatic labeling, captioning, and computer-aided detection (CAD) in radiology. Deep learning has shown superior performance in tasks such as lesion detection, classification, and reporting, potentially improving radiologists' efficiency and accuracy. However, the article also addresses limitations and considerations, including the high dependency on training data, the black box nature of deep learning, and ethical and legal issues. The conclusion emphasizes the potential of deep learning to complement radiologists by handling repetitive tasks and providing quantitative analysis, while also acknowledging the need for further development and collaboration between radiologists and computer scientists to ensure proper clinical adoption.This article provides an overview of deep learning in medical imaging, highlighting its history, development, and applications. Deep learning, a subset of machine learning (ML), has gained significant attention due to advancements in big data, enhanced computing power, and novel algorithms. The introduction of artificial neural networks (ANN) in the 1950s was limited by issues such as vanishing gradients, overfitting, and insufficient data. However, recent breakthroughs have addressed these limitations, leading to impressive performances in various fields, including medical imaging. The article discusses the evolution of ML from traditional methods to deep learning, emphasizing the importance of feature extraction and the challenges of training deep architectures. Convolutional neural networks (CNNs) and recurrent neural networks (RNNs) are highlighted for their applications in image segmentation, registration, automatic labeling, captioning, and computer-aided detection (CAD) in radiology. Deep learning has shown superior performance in tasks such as lesion detection, classification, and reporting, potentially improving radiologists' efficiency and accuracy. However, the article also addresses limitations and considerations, including the high dependency on training data, the black box nature of deep learning, and ethical and legal issues. The conclusion emphasizes the potential of deep learning to complement radiologists by handling repetitive tasks and providing quantitative analysis, while also acknowledging the need for further development and collaboration between radiologists and computer scientists to ensure proper clinical adoption.