Una revisió de neural style transfer

Abstract

[en] In recent years, a technology has emerged, capable of simulating human reasoning and performing complex tasks that conventional programming was unable to perform. We are talking about neural networks, Deep Learning algorithms that are used in many different fields and with many good results, such as: prediction of successes and simulations of all kinds, recognition and classification of elements, data processing and modeling, control engineering, artificial intelligence, facial recognition; and many more. In this work we will focus on a specific type of network, convolutional neural networks, capable of interpreting and classifying images. We will study in depth and review the problem of Style Transfer through this neural technology. Thus differentiating two phases in the work, a phase of investigation and research and a practical phase where to implement and experiment with this technology. Finally, an analysis of the operation, performance, and efficiency of this technology will be performed and an assessment will be made.