Automated detection of facial expressions using image analysis

"The broad range of applications for automatic expression detection sparks the need for a robust and effective implementation. In this paper, an exposition of the existing methods most frequently used for this purpose is done, and an analysis of their performance is carried out. These include b...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Casagrande, Lucas, Kuyumciyan, Nicolás
Otros Autores: Gambini, Juliana
Formato: Proyecto final de Grado
Lenguaje:Inglés
Publicado: 2019
Materias:
Acceso en línea:http://ri.itba.edu.ar/handle/123456789/1668
Aporte de:
Descripción
Sumario:"The broad range of applications for automatic expression detection sparks the need for a robust and effective implementation. In this paper, an exposition of the existing methods most frequently used for this purpose is done, and an analysis of their performance is carried out. These include both feature detection methods such as Gabor filters and Histograms of Gradients as well as classifiers based in neural networks. Existing data sets consisting of images of persons faces with a labeled expression are used for training and testing purposes. A success rate of 87.6% is achieved when classifying images with up to four different expressions."