Desarrollo de un sistema de detección de covid-19 para el hospital general Puyo mediante procesamiento de patrones sonoros y temperatura

In this research, a COVID-19 detection system was developed for the General Hospital Puyo using sound pattern and temperature processing. Characteristic attributes determined in recordings of cough, voice, and thermal images to differentiate healthy, flu, and COVID-19 patients through a deep learning algorithm. Sound and temperature recordings collected for each patient type, and they filtered and analyzed using Fourier transform to obtain MEL spectrograms. The thermal image of each patient analyzed with a color histogram. A response algorithm constructed using parameters of different color ranges, frequency levels, and decibels. A classification system proposed that works in conjunction with the deep learning algorithm. The system developed in Python with a low-cost computer based on Raspberry Pi-4 with microphones and thermal camera. The neural network-based algorithm was able to analyze the combination of characteristic attributes of voice, cough, and temperature patterns with a percentage that will contribute as support in COVID-19 detection. The mobile web page will present the results of the diagnosis made by the algorithm to quickly identify or rule out infected individuals without resorting to laboratory tests. Keywords: detection, COVID-19, characteristic attributes, algorithm, deep learning, Raspberry Pi-4

• Auteur principal: Luzuriaga Morán, Dayana Menaly (author)
• Format: bachelorThesis
• Langue: spa
• Publié: 2023
• Sujets: Algoritmo; Aprendizaje profundo
• Accès en ligne: http://dspace.unach.edu.ec/handle/51000/10660