DAGNet-based ANN development for image enhancement
D N Veremeev, Pavel Obukhov, K V Kislov, O Ja Kravets · AIP conference proceedings · 2021
The article is devoted to the development of a system for increasing the resolution of a raster image based on a deep artificial neural network. The article presents and describes a step-by-step plan for the development of a deep artificial neural network that solves the problem of increasing the resolution of a raster image. Examples of pairs of images from the training dataset, the developed topology of a deep artificial network, the loss function, and the training graph of the network are given. At the end of the article, an overview of the effectiveness of the developed system is provided.