Farsi Font Recognition in Document Images Using PPH Features
Yaghoub Pourasad, Houshang Hassibi, Azam Ghorbani · DergiPark (Istanbul University) · 2011
In this paper a new approach for font recognition of Farsi document images is presented. In this approach using two features PP and H, font and font size of a Farsi document image is recognized. Feature H is related to holes of imaged text document. Feature PP is related to horizontal projection profile of text lines of document image. This approach is applied on 7 widely used Farsi fonts and 7 font sizes. A dataset containing 10*49 images and another dataset including 110 images were used for testing and recognition rate more than 93.7% obtained. Images were made using paint software and were noiseless and without skew. This approach is fast and is applicable for other languages that are similar to Farsi, such as Arabic language