Gastroenterology needs its own ImageNet

Fons van der Sommen


In a period spanning less than a decade, deep learning with Convolutional Neural Networks (CNNs) has become the standard for Artificial Intelligence (AI) applications. While most of the key concepts were introduced decades ago, the two driving forces behind its success have only started building momentum at the dawn of this century: (I) an exponential increase in computational power and (II) the growing availability of large sets of labeled data. These two developments reached critical mass first in speech recognition (1), and later in computer vision, with the introduction of AlexNet (2): the first deep learning architecture to win the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) (3).