Complicated Image´s Binarization Based on Method of Maximum Variance

Author

Bai, Jie ; Yang, Yao-Quan ; Tian, Rui-li

Author_Institution

North China Electr. Power Univ.

fYear

2006

fDate

13-16 Aug. 2006

Firstpage

3782

Lastpage

3785

Abstract

Image binarization is an important preceding task in image processing and analysis. But the effect of common methods such as Ostu and Bernen are not well when the image´s luminance isn´t equal or the contrast is deficient etc. So, a binarization method based on spatial distributing and improved maximum variance is used in this paper. This method combined the characteristic of image´s spatial distributing and the statistic characteristic of maximum variance in and between clusters. The processing speed is much faster, and artifacts and broken strokes are eliminated

Keywords

image processing; pattern clustering; statistical analysis; complicated image binarization; image processing; maximum variance; spatial distribution; statistic characteristic; Analysis of variance; Character recognition; Clustering algorithms; Cybernetics; Differential equations; Histograms; Image analysis; Image processing; Machine learning; Pixel; Statistical distributions; Target recognition; Bayes Algorithm; Binarization; Maximum variance between clusters; Threshold;

fLanguage

English

Publisher

ieee

Conference_Titel

Machine Learning and Cybernetics, 2006 International Conference on

Conference_Location

Dalian, China

Print_ISBN

1-4244-0061-9

Type

conf

DOI

10.1109/ICMLC.2006.258683

Filename

4028729