Font Recognition of English Letters Using Normalized Central Moments Features

Abstract

Optical font recognition is an important process applied before or after optical character recognition. This paper presents a system for recognizing English fonts of character images. Feature selection plays a major role in recognizing the font; hence, we used normalized central moments (NCM) as the feature of choice in this study. What differentiates this study from others is the attempt to use another popular feature (distance profile features) used by other researchers and compare the results of the two. The support vector machine (SVM) method is used in training and testing. A system is developed that extracts the two features and trains two SVM models. Simulation results based on a dataset of 27,620 images belonging to three English fonts show that the proposed system can achieve an overall 94.9% correct recognition rate based on normalized central moments, while the system can achieve an overall 94.82% correct recognition rate when using distance profile features.