Supqua: aplicación móvil para interpretar datos de las señales del SITP zonal para la población en condición de discapacidad visual usando machine learning

Abstract

In this project, was develop and published a mobile application for Android devices that uses deep learning and open data features to consult the routes of the urban transport system in Bogota (SITP) on the bus stops to work offline and designed for the people in with visual disability total or partial. Was used transfer learning as a technique for getting a model able to detect objects on Tensorflow using photographs and computer-generated images, to identify the signals and for the optical character recognition of unique alphanumeric code was used Tesseract. code that is consulted in the extracted information related from the Bogota open data and available in the local storage on the device and their result is shared using the user interface with support for feedback audible to the user.