Recognition of hand written devnagari characters with percentage component regular expression matching and classification tree

Character recognition has been the subject of intensive research during the last decades. This not only because it is very challenging scientific problem but also because it provides a solution for processing large volumes of data automatically. Most of the shape representation and retrieval methods either do not well represent the shape or are difficult to do normalization (making matching hard). This paper presents new technique with the combination of encoded string and regular expressions. With Regular expressions, global shape features are captured. Characters can be in arbitrary location, scale and orientation. There are three steps in recognition which are preprocessing, feature extraction and classification. In preprocessing step character is binaries and converted into segments of consecutive ones in each column. Feature extraction is performed by representing the shape of character using regular expression Regular expression is a powerful tool and can be directly used for classification. In this paper combination of various methods along with regular expression of consonants are published .