Printed media is still popular now days society. Unfortunately, such media encountered several drawbacks. For example, this type of media consumes large storage that impact in high maintenance cost. To keep printed information more efficient and long-lasting, people usually convert it into digital format. In this paper, we built Optical Character Recognition (OCR) system to enable automatic conversion the image containing the sentence in Latin characters into digital text-shaped information. This system consists of several interrelated stages including preprocessing, segmentation, feature extraction, classifier, model and recognition. In preprocessing, the median filter is used to clarify the image from noise and the Otsu’s function is used to binarize the image. It followed by character segmentation using connected component labeling. Artificial neural network (ANN) is used for feature extraction to recognize the character. The result shows that this system enable to recognize the characters in the image whose success rate is influenced by the training of the system.
Real Time Impact Factor:
Pending
Author Name: Rio Anugrah, Ketut Bayu Yogha Bintoro
URL: View PDF
Keywords: Optical Character Recognition (OCR); segmentation; feature extraction; artificial neural network (ANN)
ISSN: 1411-8289
EISSN: 2527-9955
EOI/DOI: http://dx.doi.org/10.14203/jet
Add Citation
Views: 1
