Detección del estado fisiológico de los ojos en conductores mediante técnicas de visión artificial

• Autores: Ale Ale, Neisser, Fabián, Junior
• Formato: artículo
• Fecha de Publicación: 2019
• Institución: Universidad ESAN
• Repositorio: ESAN-Institucional
• Lenguaje: español
• OAI Identifier: oai:repositorio.esan.edu.pe:20.500.12640/3303
• Enlace del recurso: https://hdl.handle.net/20.500.12640/3303; https://doi.org/10.4067/S0718-33052019000400564
• Nivel de acceso: acceso abierto
• Materia: Fatigue detection; Artificial vision; Machine learning; HOG descriptor; CEW dataset; Detección de fatiga; Visión artificial; Descriptor HOG; https://purl.org/pe-repo/ocde/ford#2.02.04

In recent decades, the number of traffic accidents due to fatigue or drowsiness of the driver has caused significant human and material losses. At the same time, the sale in the vehicle fleet has been massified, which indicates thatpossibly in the following years, if the pertinent measures are not taken to detect fatigue, there will be an increase in automobile accidents. Therefore, in this research study, the development of a fatigue detection system in drivers that allows alerting about their status while driving using artificial vision and machine learning techniques is proposed. The techniques of these two fields of study are intercepted to generate supervised models with high performance when classifying the state of fatigue in drivers. In this study, a dataset of frontal images focusing on the physiological characteristics of the eyes was used; obtaining promising preliminary results in the detection of fatigue in real-time.