Handwritten digit recognition using classifier cooperation schemes
Journal
Proceedings of the 2nd Balkan Conference in Informatics, BCI
Date Issued
2005
Author(s)
Chakmakov, Dushan
Gjorgjevikj, Dejan
Abstract
Recent results in pattern recognition applications have shown that
SVMs (Support Vector Machines) often have superior recognition rates in
comparison to other classification methods. In this paper, the cooperation of
three SVM classifiers for handwritten digit recognition, each using different
feature family is examined. We investigate the advantages and weaknesses of
various cooperation schemes based on classifier decision fusion using statistical
reasoning. Although most of the used schemes are variations and adaptations of
existing ones, such an extensive number of cooperation schemes have not been
presented in the literature until now. The obtained results show that it is difficult to exceed the recognition rate of a single, well-tuned SVM classifier applied straightforwardly on all feature families as a single set. However, the
classifier cooperation reduces the classifier complexity and need for samples,
decreases classifier training time and sometimes improves the classifier performance.
SVMs (Support Vector Machines) often have superior recognition rates in
comparison to other classification methods. In this paper, the cooperation of
three SVM classifiers for handwritten digit recognition, each using different
feature family is examined. We investigate the advantages and weaknesses of
various cooperation schemes based on classifier decision fusion using statistical
reasoning. Although most of the used schemes are variations and adaptations of
existing ones, such an extensive number of cooperation schemes have not been
presented in the literature until now. The obtained results show that it is difficult to exceed the recognition rate of a single, well-tuned SVM classifier applied straightforwardly on all feature families as a single set. However, the
classifier cooperation reduces the classifier complexity and need for samples,
decreases classifier training time and sometimes improves the classifier performance.
File(s)![Thumbnail Image]()
Loading...
Name
Handwritten_Digit_Recognition_Using_Clas20160826-22544-n723nm-with-cover-page-v2.pdf
Size
701.96 KB
Format
Adobe PDF
Checksum
(MD5):a37bfba830b196c283b0f2d5587364aa