Mbojo Character Recognition Using Shearlet Transform and Support Vector Machine

Abstract

This paper aims to preserve one of the Indonesian culture, Mbojo Character. Mbojo character recognition system will be created by utilizing pattern recognition using Optical Character Recognition (OCR) technique with Shearlet Transform method for feature extraction and Support Vector Machine (SVM) for classification. Data used in this study is the image of mbojo words that consist of 2 characters typed using bimambojo.otf font with the size of 9pts for each word and with image size of 50×50 pixels. The training data uses 150 word images which represents all of Mbojo characters where each word has 3 images with 3 different positions of character placement, such as above, in the middle, and below the image field. While the testing data uses 50 word images where each of the character placed randomly in the image. All of the data were preprocessed by using grayscaling, binarization, and centering regions methods. The implementation of centering region method makes the system able to achieve accuracy up to 90 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> .

Access to Document

Other files and links

• Link to publication in Scopus

Fingerprint