Sistema interactivo de monitoreo y reconocimiento emocional basado en IA para niños con TEA

Abstract

CONDITIONING OF PROJECT PUBLICATION. Many children with Autism Spectrum Disorder (ASD) experience difficulties in expressing their emotions and understanding their surroundings, which can complicate their therapeutic processes and daily activities. In response to this situation, the present project proposes the design of a system that allows the identification of a child’s emotional state while playing, with the aim of providing specialists with more accurate information to support decision-making during interventions. This proposal is justified by the need for accessible technological tools that contribute to more appropriate and personalized support. For the development of the project, a Raspberry Pi 3 was used as the main processing unit, along with a sensor-based controller built using an ESP32. In addition, the system was integrated with the Google Cloud artificial intelligence platform (Vertex AI) for facial recognition and emotion identification. The tests were carried out with children between 5 and 10 years of age, following technical guidelines and under the supervision of a therapist. The results obtained showed favorable performance, achieving an 87 percent accuracy rate in detecting the emotion of happiness, which demonstrated the feasibility of the system in a real environment. In conclusion, this project shows that low-cost technology can be applied to better understand the emotions of children with ASD and to strengthen the support provided in educational and therapeutic settings. Keywords:Autism Spectrum Disorder, emotion recognition, educational technology, therapeutic support.