Radiologie Apr 2026

In radiology, are complex numerical descriptors automatically extracted from medical images using deep learning architectures, specifically Convolutional Neural Networks (CNNs) . Unlike traditional "handcrafted" radiomic features that represent human-defined traits like size or texture, deep features are learned by the AI model during training, often capturing abstract patterns that may be invisible to the human eye. Key Characteristics of Deep Features