Sistema de reconocimiento facial basado en imágenes con color

This paper develops an algorithm system to check whether the role of color can be an important attribute in facial recognition systems in two dimensions (2-D), with frontal orientation and small variations in the gestures of individuals. The first phase involves the detection and localization of the...

Descripción completa

Detalles Bibliográficos
Autores: Pedraza Pico, Beatriz Omaira, Rondón, Paola, Arguello, Henry
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2011
País:Colombia
Institución:Universidad Industrial de Santander
Repositorio:Repositorio UIS
Idioma:español
OAI Identifier:oai:noesis.uis.edu.co:20.500.14071/8216
Acceso en línea:https://revistas.uis.edu.co/index.php/revistauisingenierias/article/view/113-122
https://noesis.uis.edu.co/handle/20.500.14071/8216
Access Level:acceso abierto
Palabra clave:Principal Component Analysis (PCA)
eigenfaces
AdaBoost
euclidean distance
mahalanobis distance
Análisis de Componentes Principales (PCA)
distancia euclidiana
distancia mahalanobis
Descripción
Sumario:This paper develops an algorithm system to check whether the role of color can be an important attribute in facial recognition systems in two dimensions (2-D), with frontal orientation and small variations in the gestures of individuals. The first phase involves the detection and localization of the human face for which the learning algorithm uses a combination of AdaBoost and cascade classifiers to increase detection rates. In a second phase the eigenfaces approach is applied and a clasification system is implemented, to recognize and identify the subject of entry to a specific individual, using the Euclidean and Mahalanobis distance. We illustrate the results of the proposed system for both color images as gray, finding that the color information at the HSV plane can improve recognition rates when compared with the RGB plane.