Program SVM :

Machine Learner to Recognize Arabic Text in Images

Program Abilities :

• The code builds its own database by extraxting single words from .pdf , .doc files , or by reading an input cell array of characters

• The code generates 127 sample image of each word with different fonts and sizes

• Extracts a histogram of oriented gradient features for image regiteration

• Trains a multi-class support vector machine classifier

• Reads jpg , jpeg , png , bmp , tif , tiff image files

• Image segmentation technique to extract single words from images

• Writes the output text in a separate .txt file