Resumen

RAMIREZ CABEZAS, Jhensy Alexander; JIMENEZ BEDON, Alexi Marcelo; BAJANA ZAJIA, Johnny Xavier  y  CUNUHAY CUCHIPE, Wilmer Clemente. Mobile application for communication between listeners and people with hearing disabilities through phrase recognition in LSEC with convolutional neural networks. e-Rev. Multidiscip. Saber [online]. 2025, vol.3, e-RMS04122025.  Epub 17-Mar-2026. ISSN 2960-2467.  https://doi.org/10.61286/e-rms.v3i.307.

Communication constitutes a fundamental human right and a determining factor for social inclusion. In Ecuador, people with hearing disabilities face structural, linguistic, and technological barriers that limit their full interaction in education, work, and healthcare settings. The officially recognized Ecuadorian Sign Language (ESL) lacks robust technological solutions that facilitate real-time translation. This study presents the development of an inclusive mobile application for two-way communication of 15 sentences in LSEC, using convolutional neural networks (CNNs) optimized with TensorFlow Lite to ensure its operation on low-cost mobile devices. A proprietary dataset was created with visual records, developed with the participation of the deaf community, ensuring cultural and linguistic relevance. The applied methodology combined a mixed approach: technical validation of the model using accuracy, F1-score, and latency metrics, along with evaluation of usability and social acceptance in field tests with 354 statistically selected participants. The results showed an accuracy rate of over 90%, an average response time of 0.8 seconds per sentence, and broad social acceptance, confirming that the app mitigates communication asymmetry and strengthens the autonomy of the deaf community, highlighting its usefulness in educational and healthcare settings. This pioneering contribution to digital accessibility in Ecuador and Latin America, offering a replicable and scalable solution for underrepresented sign languages. Furthermore, it aligns with Sustainable Development Goals 4, 9, and 10, strengthening its scientific, technological, and social relevance.

Palabras clave : digital accessibility; hearing impairment; artificial intelligence; ecuadorian sign language; convolutional neural networks.

        · resumen en Español     · texto en Español     · Español ( pdf )