DocumentCode :
2005334
Title :
Color pixels classification in an hybrid color space
Author :
Vandenbroucke, N. ; Macaire, L. ; Postaire, J.-G.
Author_Institution :
Ecole des Mines de Douai, France
Volume :
1
fYear :
1998
fDate :
4-7 Oct 1998
Firstpage :
176
Abstract :
We describe a new approach for color image segmentation which is considered as a problem of pixels classification. The algorithm is applied to classify pixels of soccer color images in order to recognize the team of players. The color pixels classification algorithm is achieved in a new color space, the hybrid color space, which is determined thanks to a supervised learning scheme. This space is defined by the set of the three most discriminating color features selected among a set of classical color features
Keywords :
feature extraction; image classification; image colour analysis; image recognition; image segmentation; learning (artificial intelligence); algorithm; color features; color image segmentation; color pixels classification; hybrid color space; soccer color images; soccer player recognition; supervised learning; Classification algorithms; Image analysis; Image color analysis; Image recognition; Image resolution; Image segmentation; Labeling; Pixel; Space technology; Supervised learning;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Image Processing, 1998. ICIP 98. Proceedings. 1998 International Conference on
Conference_Location :
Chicago, IL
Print_ISBN :
0-8186-8821-1
Type :
conf
DOI :
10.1109/ICIP.1998.723452
Filename :
723452
Link To Document :
بازگشت