A New Method for Identification of Partially Similar Indian Scripts

In this paper, the texture symmetry/non-symmetry factor has been exploited to identify the Indianscripts. Biwavelants have been proposed to obtain the script texture using third order cumulantand bispectra. As the Indian scripts are partially similar to each other, in order to identify them,the samples must include more number of dissimilar characters. The features of individual linesare added repeatedly to enhance the dissimilarity until it reaches to a saturation level which inturn is used to compute a confidence factor i.e. amount of confidence attained in identifying aparticular script sample. This variation in confidence factor also gives an estimate of the optimumsample size (number of lines) required for expected results. Cumulants are sensitive to the scriptcurvatures and therefore are most suitable for the partially similar Indian scripts. The doublediscrete Fourier transform of third order cumulant gives bispectra which estimates the factor ofsymmetry/non-symmetry in terms of the quadratically coupled frequencies. The envelope ofbispectra (biwavelant) obtained using wavelet (db8) provides an accurate behavior of the scripttexture; which along with Newton-Raphson technique is used to classify the Indian scripts.Various classifiers have been tested for script identification and out of them SVM gives the bestresults. The method successfully identified the 8 Indian scripts like Devanagari, Urdu, Gujarati,Telugu, Assamese, Gurmukhi, Kannada, and Bangla with desired accuracy.