Sundanese Aksara Recognition Using Histogram of Oriented Gradients

Abstract

Indonesia is a famous nation for its wealth in both natural and language resources and culture. Aksara is one of the Indonesian cultures that must be preserved therefore, as not to lose its existence. To avoid the loss of the existence of letters, especially Sundanese aksara, we proposed a new approach Sundanese word recognition with considering rarangkèn characteristic using the Histogram of Oriented Gradients method and support vector machine as a classification method. The datasets used are sourced from a Sundanese dictionary book. Based on the test results obtained an accuracy 81.48 % of the recognition of word Sundanese aksara with the values pixels per cell is 10x10 and cells per block is 1x1 or the values pixels per cell is 20x20 and cells per block is 3x3.