Embedded system for the recognition of writing on embedded Linux

Abstract

In this article, an embedded system for handwriting recognition or characters using artificial neural networks    is presented. The type of RNA that is used is the self-organizing map (SOM Self Organizing Map) because it consumes less memory and it is low computational cost. The hardware development board used was the BEA GLEBONE black connected to a resistive touch screen. The card has a black BEAGLEBONE ARM Cortex-A8 processor and has as Linux embedded operating system. A graphical interface using the Qt library for embedded linux. In this interface, characters were entered using a touch pencil and then extract features from the image and process them. The implemented system was tested as part of a calculator application that performs the four basic operations, and finally, the evaluation of the performance of self-organizing map (SOM) was held.