Identification of MPEG-4 FDP patterns in human faces using data-mining techniques

Date issued

2005

Journal Title

Journal ISSN

Volume Title

Publisher

Václav Skala - UNION Agency

Abstract

In this paper we try to induce rules that describe patterns in human faces. We apply two different data-mining algorithms, C4.5 and C5.0, in a database of faces parameters in the MPEG4 FDP (Face Definition Parameters) form. Also we modify the database in two different ways before applying the algorithms: variable discretization of some fields; and selection of the main clusters with self-organizing maps.

Description

Subject(s)

data mining, samoorganizované mapy

Citation

WSCG '2005: Posters: The 13-th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision 2005 in co-operation with EUROGRAPHICS, University of West Bohemia, Plzen, Czech Republic, p. 9-10.
OPEN License Selector