Title :
A Writer Recognition approach Based on SVM
Author :
Yin, Zhonghang ; Yin, Peipei ; Sun, Fuchun ; Wu, Haomiao
Author_Institution :
Dept. of Comput. Sci. & Technol., Tsinghua Univ., Beijing
Abstract :
A writer recognition approach based on SVM is proposed in this paper. Above all, the effect of SVM on the writer recognition is also analyzed under the condition of small samples and weak features, and then the improvement of its performance is validated by experiments. Futhermore, in terms of C-SVC, two parameter optimization schemes are put up based on collapse tests plus margin-based bounds and collapse tests plus compression bounds, respectively, their performance are compared, thus the generation performance of the classifier is enhanced
Keywords :
image recognition; optimisation; pattern classification; support vector machines; classifier; collapse tests plus compression bounds; collapse tests plus margin-based bounds; parameter optimization; support vector machines; writer recognition; Application software; Computer science; Feature extraction; Image segmentation; Intelligent systems; Sun; Support vector machine classification; Support vector machines; Systems engineering and theory; Testing; C-SVC; writer identification;
Conference_Titel :
Computational Engineering in Systems Applications, IMACS Multiconference on
Conference_Location :
Beijing
Print_ISBN :
7-302-13922-9
Electronic_ISBN :
7-900718-14-1
DOI :
10.1109/CESA.2006.4281720
