B. Harijanto, E.L. Amalia, M. Mentari
Maps are one form of an image that often encountered in various interests. For example, many books on tourist attractions or other information that provide maps as a media of information. However, sometimes people with visual impairments such as presbyopia, hypermetropia, or astigmatism have difficulty reading the map which is usually given in small size, multi orientation (multi-scale and multi-direction). Therefore, this study tries to provide a solution through an application that converts image to text in image conditions that have many orientations and different things are briefly called heterogeneous text. The Optical Character Recognition (OCR) system that was built beginning with taking pictures made through a cell phone camera as the first step in obtaining a digital map file, then enters the pre-processing, text segmentation, feature extraction from each different character, then continues to the classification stage. This OCR system for recognizing text with multiple orientations will help people make digital maps easier to read, especially for people who have presbyopia, hypermetropia, or astigmatism vision problems. The proposed model achieves good average accuracy for classifying the characters in various orientation successfully. © Published under licence by IOP Publishing Ltd.
Dept. of Informatics Engineering, State Polytechnic of Malang, Malang, Indonesia