Multi-font Numerals Recognition for Urdu Script based Languages
Imran Razzak, Syed Afaq Hussain, Abdel Belaı̈d, Muhammad Sher · 2009
Abstract—Handwritten character recognition of Urdu script based languages is one of the most difficult task due to complexities of the script. Urdu script based languages has not received much attestation even it is used more than 1/6th of the population. The problem in handwritten numeral recognition is the shape similarity between handwritten numerals and dual style for Urdu script. This paper presents a fuzzy rule base, HMM and Hybrid approaches for the recognition of numerals both Urdu and Arabic in unconstrained environment from both online and offline domain for online input. Basically offline domain is used for preprocessing i.e normalization, slant normalization. The proposed system is tested and provides accuracy of 97.1%.