Self organizing topological tree for skin color detection

Author

Deng, Xiaoyun ; Xu, Pengfei ; Chang, Chip-Hong

Author_Institution

Center for High Performance Embedded Syst., Nanyang Technol. Univ., Singapore

Volume

2

fYear

2004

fDate

6-9 Dec. 2004

Firstpage

1097

Abstract

Self organizing topological tree (SOTT) evolved from self organizing map (SOM) has advanced features that make it suitable for online clustering. It can be used in applications such as image compression, data mining, and pattern identification. This work in particular explores the use of SOTT in skin color detection. Three different schemes have been proposed and evaluated in terms of their performance, speed, memory requirement, and the size of the training set. Performances of SOTT as under these three schemes have also been benchmarked and compared against other important skin color classification methods.

Keywords

computational complexity; image colour analysis; object detection; pattern clustering; self-organising feature maps; topology; trees (mathematics); unsupervised learning; computational complexity; data mining; image compression; online clustering; pattern identification; self organizing map; self organizing topological tree; skin color classification methods; skin color detection; training; Bayesian methods; Embedded system; Face detection; Human computer interaction; Network topology; Neural networks; Neurons; Organizing; Skin; Training data;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems, 2004. Proceedings. The 2004 IEEE Asia-Pacific Conference on

Print_ISBN

0-7803-8660-4

Type

conf

DOI

10.1109/APCCAS.2004.1413075

Filename

1413075