Optical Character and Font Recognizer

Abstract

Optical Character and Font Recognizer focuses primarily on building a complete model for document processing. The proposed system recognizes the font style along with the text from an image of certain resolution. The system uses principles of both machine learning and image processing to obtain the desired results. The model uses Contour selection for character extraction and K-Nearest Neighbor approach for character and font recognition. With the assistance of the proposed system using the mentioned techniques, scanned documents can be altered or the font style of a particular document can be known as desired. Many models that perform character recognition are present but a model that performs both character and font recognition with good accuracy is difficult to find. The experiment resulted in 87% overall accuracy for detection of characters.