Return to the institutional site
    • español
    • English
View Item 
  •   Séneca Home
  • TRABAJOS DE GRADO
  • Maestría
  • View Item
    • español
    • English
  •   Séneca Home
  • TRABAJOS DE GRADO
  • Maestría
  • View Item
JavaScript is disabled for your browser. Some features of this site may not work without it.

Browse

All of SénecaCommunities and CollectionsFaculties and ProgramsAuthorsTitlesSubjectsTypes of contentsAuthor profilesThis CollectionFaculties and ProgramsAuthorsTitlesSubjectsTypes of contents

My Account

LoginRegister

Statistics

View Usage Statistics

Information of interest

What is SénecaHow publishGuidelinesContact us

Facial action unit detection with convolutional neural networks

RISMendeley
URI: http://hdl.handle.net/1992/13826
Author: Romero Vergara, Andrés FelipeUniandes authority
Director(s)/Advisor(s): Arbeláez Escalante, Pablo AndrésUniandes authority
Publication date: 2017
Content type: masterThesis
Keywords:
Redes neurales (Computadores) - Investigaciones
Procesamiento de imágenes - Investigaciones
Sistemas de reconocimiento de configuraciones - Investigaciones
Expresión facial - Procesamiento de imágenes - Investigaciones
Abstract:
We propose a novel deep convolutional neural network architecture to study the problem of action unit detection. We leverage recent gains in large-scale object recognition by formulating the task of predicting the presence of a specific action unit in a still image as simple image-level binary classification. We first train a convolutional encoder on the problem of multi-view emotion recognition as a high-level representation of facial expressions. We show that our architecture generalizes across views, ethnicity, gender and age by merging and training jointly on three standard emotion recognition datasets: CK+, Bosphorus and RafD. Our system is the first fully multi-view emotion recognizer proposed in the literature. We then extend this shared learned representation with fully-connected layers trained to detect individual action units. Our approach is conceptually simpler and yet significantly more accurate than the best methods based on the dominant paradigm for the study of this problem, which relies on facial landmark detection as an intermediate task. We conduct experiments on the BP4D dataset, the largest and most challenging benchmark currently available for action unit detection, and report an absolute improvement of 16% over the previous state-of-the-art
Show full item record

Files in this item

Thumbnail
Name:
u729464.pdf
Size:
28.55Mb
Format:
PDF

Statistics

View Usage Statistics
Donaciones

Los Andes

Donaciones


Icono Repositorio

Los Andes

Repositorio


Icono Egresados

Los Andes

Egresados


Icono Eventos

Los Andes

Eventos



Cra 1 Nº 18A - 12

Bogotá - Colombia

Postal code: 111711

+(571) 339 49 99

+(571) 339 49 49


Normatividad Institucional

  • Actos internos e incremento
  • Bienestar
  • Derechos pecunarios
  • Estatuto docente
  • Estatuto general
  • Ley de transparencia
  • Porcentaje de incremento
  • Reglamentos de estudiantes
  • Uso de datos personales

Enlaces Rápidos

  • ATC (Acceso Temporal al Campus)
  • Universidad de los Andes Caribe
  • Convivencia y transparencia
  • Educación Continuada
  • Emergencias: Extensión 0000
  • Nuestros profesores
  • Mapa del sitio
  • Multimedia
  • Noticias
  • Preguntas frecuentes

Redes sociales

  • Facebook
  • twitter
  • youtube
  • linkedin
  • instagram
  • snapchat
  • vimeo
  • google

Directorio de redes