Abstract:

This paper presents the development and evaluation of a multi-platform communication application designed for children with nonverbal autism. Utilizing Python and deep learning models, the application, named Conectando Voces, aims to facilitate communication for children with autism by providing a user-friendly interface and voice generation capabilities. The application employs the Flet framework for building the UI, enabling deployment across various devices including mobile phones, tablets, and PCs. Voice generation is achieved through Google’s text-to-speech model, with a fallback to pyttsx3 for offline use. The validation study involved five children with varying degrees of autism and comorbidities, highlighting the app’s potential effectiveness in promoting speech. Results indicate that children with mild to moderate autism benefit significantly, while those with severe autism or additional comorbidities require further tailored support. The study underscores the importance of personalized adaptations and comprehensive support for maximizing the app’s utility in enhancing communication skills in children with autism.