Identification of Kashmiri Script in a Document Image Using Dwt and Entropy

Abstract

Over the past decades, the growth in the development of automated systems and their related technologies for document processing has been exponential to achieve more and more effective & efficient solutions. Since the majority of documents contain text so text processing is an important aspect of the document processing. However for text processing it is important to know in which script a particular document is written. These scripts are used to write languages. The identification of script has been an important area of research and accordingly a substantial amount of research work has been done using different schemes. However, there is very less work reported with respect to the identification of Kashmiri language. The main impetus in this work is to perform the identification of Kashmiri Script alongside Urdu, Devanagari and English. The scheme uses discrete wavelet transformation and entropy evaluation. The results achieved are promising. This work will also pave a way for designing more optimal and better solutions with respect to overall automated solutions to the Kashmiri Language.