Recognition of Ancient stone Inscription Characters Using Histogram of Oriented Gradients
5 Pages Posted: 7 Aug 2019
Date Written: August 5, 2019
Abstract
The transliteration of ancient stone inscription characters into present characters is required in the area of epigraphy. Optical Character Recognition is not adopted as the rock surface becomes texture of an image posing several challenges like degradation due to weather conditions and minimal distinction between foreground and background. In this paper, a new feature using Histogram of Oriented Gradients is proposed for ancient stone inscribed character recognition that is independent to any language script. This feature counts occurrences of gradient orientation in localized portions of an image. Here 11th century stone inscription digital images are used for creating database. Once the features are computed, the system is trained up with Support Vector Machine for classifying them into modern characters. Experiments are performed on 10 characters of 300 database size and showed significant recognition rate.
Keywords: Character recognition, Ancient stone inscriptions, Feature Extraction, Histogram of oriented gradients, Support vector machine
Suggested Citation: Suggested Citation