Face-Reco. Diseño de un sistema para detección y análisis de rostros

This project arises as a response to the current need to incorporate face recognition and analysis functionalities in the different software applications. During the execution of the project, a system was built that could detect and analyze faces. To analyze means that the system can determine the gender of the faces identified in the image and can determine an emotion according to the model of 6 basic emotions proposed by Ekman: joy, sadness, anger, surprise, disgust, fear, and a neutral state as the absence of emotion.In this document we formulated a project oriented to the detection and analysis of faces which developed an architecture oriented to web services, so that other developers incorporate functionalities of detection and analysis of faces can use in their systems.According to the results obtained, the greater precision of the system is associated with the emotion of happiness. The masculine gender presented greater errors in classification than the feminine gender. On the other hand, with the emotions of anger and fear, the performance of the system, in terms of classification power, decays.One of the areas in which a significant and growing use of this type of system is observed is related to mobile devices and virtual humans.

• Autores: Girón Arias, Andrés Felipe; Chacón Peña, Juan Pablo
• Fecha de publicación: 2019
• Institución: Universidad de San Buenaventura
• Repositorio: Repositorio USB
• Idioma: spa
• OAI Identifier: oai:bibliotecadigital.usb.edu.co:10819/7440
• Acceso en línea: http://hdl.handle.net/10819/7440
• Palabra clave: Visión por computador; Redes neuronales; Python; Género; Emociones; Aprendizaje de máquina; Computer Vision; Neuronal Networks; Gender; Emotions; Machine Learning; Multimedia interactiva; Diseño de sistemas; Arquitectura de software
• License: Atribución-NoComercial-SinDerivadas 2.5 Colombia