Authors:
Adriano Martins Moutinho
1
;
Antonio Carlos Gay Thomé
1
;
Luiz Biondi Neto
2
and
Pedro Henrique Gouvêa Coelho
2
Affiliations:
1
Núcleo de Computação Eletrônica, Universidade Federal do Rio de Janeiro, Brazil
;
2
Faculdade de Engenharia, Universidade do Estado do Rio de Janeiro, Brazil
Keyword(s):
Face detection, neural networks, image processing.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Artificial Intelligence and Decision Support Systems
;
Biomedical Engineering
;
Biomedical Signal Processing
;
Computational Intelligence
;
Enterprise Information Systems
;
Health Engineering and Technology Applications
;
Human-Computer Interaction
;
Methodologies and Methods
;
Neural Network Software and Applications
;
Neural Networks
;
Neurocomputing
;
Neurotechnology, Electronics and Informatics
;
Pattern Recognition
;
Physiological Computing Systems
;
Sensor Networks
;
Signal Processing
;
Soft Computing
;
Theory and Methods
;
Verification and Validation of Knowledge-Based Systems
Abstract:
Security systems based on face recognition often have to deal with the problem of finding and segmenting the region of the face, containing nose, mouth and eyes, from the rest of the objects in the image. Finding the right position of a face is a part of any automatic identity recognition system, and it is, by itself, a very complex problem to solve, normally being handled separately. This paper describes an approach, using artificial neural networks (ANN), to find the correct position and separate the face from the background. In order to accomplish this goal, a windowing method was created and combined with several image pre-processing steps, from histogram equalization to illumination correction, as an attempt to improve neural network recognition capability. This paper also proposes methods to segment facial features such as mouth, nose and eyes. Finally, the system is tested using 400 images and the performance of face and facial features segmentation is presented.