@inproceedings{1d16a50fab494aa0a50ef59216200582,
title = "Human face identification using invariant descriptions and a genetic algorithm",
abstract = "A new method to automatically identify a human face onto a 2D gray level image is presented. The method uses an invariant description of the face and a genetic algorithm to accomplish the task. The features used are the first four translation, rotation and scale moment invariants proposed by Hu [1]. In a first step, an image possibly containing a face is first divided into small cells of fixed size of 5 × 5 pixels. For each cell, the ordinary moments are next computed. From these, the corresponding Hu{\textquoteright}s invariants are then derived. Human face identification is thus accomplished by grouping individual cells using a genetic algorithm by fitting a specific cost function. This cost function corresponds to the invariant description of a human face given in terms of the detected image features.",
author = "R. Pinto-El{\'i}as and Sossa-Azuela, {J. H.}",
note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 1998.; 6th Ibero-American Congress on Artificial Intelligence, IBERAMIA 1998 ; Conference date: 05-10-1998 Through 09-10-1998",
year = "1998",
doi = "10.1007/3-540-49795-1_26",
language = "Ingl{\'e}s",
isbn = "3540649921",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "293--302",
editor = "Helder Coelho",
booktitle = "Progress in Artificial Intelligence - IBERAMIA 1998 - 6th Ibero-American Conference on AI, Proceedings",
address = "Alemania",
}