Regresar

Intelligent System for the Learning of Sign Language Based on Artificial Neural Networks

Abstract:

Sign language has become a form of communication for people who have some type of hearing impairment, in order to relate to the world around him and perform daily activities including education, work performance and personal development. In the knowledge society, the development of new Technologies for Information and Communication (TIC), they constitute a teaching tool that changes the conditions for learning and communication as the world is digitized. Therefore, the development of an Ecuadorian sign language learning system based on the use of a gesture sensor and an artificial neural network multilayer feed-forward, implementing the Backpropagation algorithm, which takes advantage of the parallel property of decreasing the time required by a processor to distinguish the existing relationship between given patterns, represents a solution for improving the teaching-learning process. The system consists of three main layers: The first is responsible for the acquisition of data through a gestural device; the second uses the library of the gesture sensor to perform the acquisition and storage of information of the hand and the position of the fingers. Finally, the output layer, once the training of the neuron is analyzed, generates the real-time recognition of sign language.