Arabic Font Recognition Based on a Texture Analysis

摘要

Existing works on the font recognition and based on texture analysis often used Gray Level Cooccurence Matrix (GLCM), Gabor Filters (GF) and wavelet. In this paper, we use Steer able Pyramid (SP) for texture analysis of arabic homogeneous and normalized text block in order to font recognition. In this frameworks, we use K Nearest Neighbors (KNN) and Back-propagation Artificial Neural Network (BpANN) for classification. The Obtained experimental results on the APTID/MF database (Arabic Printed Text Image/ Multi-Font) are encouragents.