سال انتشار: ۱۳۹۰

محل انتشار: هفتمین کنفرانس ماشین بینایی و پردازش تصویر

تعداد صفحات: ۴

نویسنده(ها):

Akram Alsadat Hajiannezhad – Electrical and Computer Engineering DepartmentSemnan UniversitySemnan, Iran
Saeed Mozaffari – Electrical and Computer Engineering DepartmentSemnan UniversitySemnan, Iran

چکیده:

In this paper, a new method based on fractal geometryis proposed for Farsi/Arabic font recognition. The featureextraction does not depend on the document contents whichconsiders font recognition problem as texture identification task.The main features are obtained by combining the BCD, DCD,and DLA techniques. Dataset includes 2000 samples of 10typefaces, each containing four sizes. The average recognitionrates obtained for these 10 fonts and 4 sizes (40 classes) usingRBF and KNN classifiers are 96% and 91% respectively. Thedimension of feature vectors extracted by the proposed fractalapproach is very low. This property obviates the need fornumerous training samples. Experimental results show that thisalgorithm is robust against skew. Simultaneously identifying typeand size of the font is the most important innovation of thispaper.