Resumen

GONZALEZ-ARRIAGA, Daniel Marcelo; VARGAS-TREVINO, María Aurora Diozcora; GUERRERO GARCIA, Josefina  y  LOPEZ GOMEZ, Jesús. Development of a Platform for Generation of CNN and Multilayer Neural Networks. Comp. y Sist. [online]. 2022, vol.26, n.1, pp.171-182.  Epub 08-Ago-2022. ISSN 2007-9737.  https://doi.org/10.13053/cys-26-1-4162.

This research presents the design of a platform that assists in the generation of convolutional (CNN) and multilayer neural networks to provide a user-friendly interface for the design, formation, and development of neural networks. This platform is developed in LabVIEW as this software allows to perform inter-faces and generate an executable for use. It aims to reduce the development time of neural networks by providing an assistant-like graphical interface that guides the user through various common scenarios (data import, neural network construction and adjustment), allows the user to focus on solving their problems without having to write code, edit text files, or manually analyze recorded data. The user interface with the options offered is described. The way the neural network is generated is described. The results generated with the platform are presented producing an image with the proposed methodology applying a complete convolution layer. The usefulness of this platform is explained by presenting a case where there is a significant improvement in the development of a neural network, in time and reduction of errors.

Palabras llave : CNN; multilayer; software; platform; LabVIEW platform; MATLAB platform.