Feature Extraction in Kawi Script

Abstract

The Kawi script is a derivative of the post-palawa language. Kawi itself in Sanskrit means poet. The Kawi script itself is found in many ancient manuscripts from ancient times. Kawi script itself nowadays is no longer used, many people don't know Kawi script. In this modern era, where everything is digital, it needs preservation, one of which is by using computers to recognize kawi script patterns. Before identifying characters, it is necessary to have digital image information, one of which is the extraction process. This research will create a feature extraction system for the kawi script which will later be used as input for the classification of the kawi script. This study uses data sourced from books and in this research, the data taken is only 6 types of data. In the process of making this system using the Matleb application. In the testing phase, the GLCM (Gray Level Co-Occurrence Matrix) feature extraction will be used which includes Contrast, Correlation, Energy, and Homogeneity, then identification will be processed. The results of this study produce values ​​from the GLCM method, namely values ​​from Contrast, Correlation, Energy, and Homogeneity. It is expected that the values ​​of the 4 features can be used as input data from the classification from  in further research.