Diseño de un sistema de interacción mediante visión artificial para personas con discapacidad motora y dificultad del habla.

Abstract

In the present study a system was developed focused in people with moto disability and speech absence that still preserve the voluntary movement of face muscles like the visual detection, this device allows the selection of images in a monitor using vision techniques by computer. The artificial vision process for this system is based on the identification of eyes movement, for which an extraction of the characteristics is done to identify the face, later the eyes, recognition of pupils’ movements and finally make an automatic recognition of these patterns. The system integrates a graphical interface thas consists of a main menu of images referring to six grammatical categories, each category displays a submenu that includes several images according to their classification. The computer screen was devided into three parts, each one with a specific image, the camera located at the top of the monitor captures the direction of the users gaze to identify the image that is being observated. The objective is selects by fixing the gaze at that point for a specific time interval, once selected a category can be navigated trough each by a short blink, as a longer blink will allow go back to the main menu, the algorithm of the system was developed in Python due to its diversity of library applications, and the OPENCV library was used for the variaty of commands that allows to use for image processing. The graphic user interface was made in QT Creator due to the flexibility of its integration with Python, all these are open source platforms which permits the system to be very reasonably priced and have greater social reach.