Implementación de un sistema con inteligencia computacional para identificar dificultad respiratoria a partir del procesamiento digital de señales de voz

Abstract

The developed system incorporates computational intelligence, this allows to identify people with respiratory distress (caused by influenza) in an automatic and non-invasive way, from the digital processing of voice signals, integrating the calculation of acoustic, spectral, temporal and statistical parameters, which implemented in a client-server architecture, they allow the respective analysis, obtained as a result of the recording of two sustained vowel sounds (vowel "a" and "o"), through the microphone of a mobile device of a Spanish-speaking population group aged between 18 and 49 years old. The results obtained determine that for the detection of respiratory distress, the vowel "a" is more efficient in women with a hit rate of 97.62% and the vowel "o" in men; reaching a hit rate of 96.77%. This system is expected to contribute to new support tools with potential for application in health, to promote biosafety protocols, especially in this context of a Covid-19 pandemic; illness that causes respiratory distress, like the flu.