Face Fitting using a Genetic Algorithm.
Abstract
The ability to estimate the three-dimensional structure of a human face from a single image has many useful applications, including but not limited to the fields of face recognition, modelling and actors head replacement. Although a lot of promising techniques have been developed in this area a number of significant technical hurdles exist preventing the development of a reliable method to extract three-dimensional structure without human intervention. In this poster we present work in progress on a new technique that should overcome some of the problems associated with the current state-of-the-art methods. We intend to use Genetic Algorithms to search the face space to find a match with a given input image. We hope to show that this method will avoid local-minima problems associated with traditional gradient descent techniques as well as provide a method that locates the position, pose and shape of the face without any user interaction.
BibTeX
@inproceedings {10.2312:egp.20101031,
booktitle = {Eurographics 2010 - Posters},
editor = {Anders Hast and Ivan Viola},
title = {{Face Fitting using a Genetic Algorithm.}},
author = {Hunter, D.W. and Tiddeman, B. P. and Perrett, D. I.},
year = {2010},
publisher = {The Eurographics Association},
DOI = {10.2312/egp.20101031}
}
booktitle = {Eurographics 2010 - Posters},
editor = {Anders Hast and Ivan Viola},
title = {{Face Fitting using a Genetic Algorithm.}},
author = {Hunter, D.W. and Tiddeman, B. P. and Perrett, D. I.},
year = {2010},
publisher = {The Eurographics Association},
DOI = {10.2312/egp.20101031}
}