Character recognition is becoming extremely important in the present day in office automation and in¬telligent system development Schemes are available for printed character analysis and recognition in different scripts world wide Successful attempt shave been made to wards the achievement of this goal in few In¬dian sripts. However, commercial products are still a rarity. Here an attempt to develop such a system, for recognition of printed Odiya charaters, is taken and success to some extent is reported.
Various pre-processing and segmentation unit are described in brief. Units are described in brief. Necessary technical details of implementation is also provided, wherever necessary. Second order geometric moments, after dividing the character in to a number of rectangular sectors, are extracted for representation of these characters. Nearest neighbour (NN) and k-NN classifiers are utilized as decision making devices. Results obtained from different feature classifier combinations are also discussed.
Added on September 23, 2014
Product Type : Research Paper
License Type : Freeware
System Requirement :
Author : Peeta Basa Pati, A G Ramakrishnan, U K Arvind Rao