Desarrollo de un kit educativo basado en visión por computador y redes neuronales de aprendizaje profundo para la identificación de campos léxicos, procesos de prelectura y lectura en el soporte educativo de niños de primero de básica.

Abstract

An interactive educational system was created for first-grade children in Ecuador. It uses a YOLOv8 model to detect pictograms related to lexical fields. The system provides real-time feedback through text-to-speech, helping children correctly identify and classify the pictograms. The implementation was evaluated with children, yielding positive results in terms of usability.