Authors : Natalie Segura Velandia, Ruben D. Hernandez Beleno and Robinson Jimenez Moreno
Abstract: This study presents a review of the computational methods implemented with the deep learning technique, highlighting the architecture of convolutional neural networks. Works that use these methods are addressed for image recognition and pattern recognition, each characterizing a specific task. With the technological and investigative momentum that has been generated today and with the development of systems with the capacity to process data with advanced intelligence equivalent to that of the human being, it has awakened an awareness of technological applications in all sectors from business to health where data processing takes place at times less than the human brain could do. In this way, deep learning is an approximation to human perception based on the hierarchical functioning of the neocortex, a fundamental part of the human brain. This technique relies on computational models composed of several layers of processing to learn the representations of data with several levels of abstraction, possessing the ability to modify its internal parameters which are used to calculate the representation in each layer from the previous layer.
Natalie Segura Velandia, Ruben D. Hernandez Beleno and Robinson Jimenez Moreno, 2017. Applications of Deep Neural Networks. International Journal of Signal System Control and Engineering Application, 10: 61-76.