Towards a Font Classification Model for Romanian Cyrillic Documents

Abstract

This paper presents a solution on how to classify the fonts in the 17th century Romanian Cyrillic documents. This solution is based on a mix of unsupervised and supervised machine learning technics. The unsupervised process is the application of K-Means method to create the dataset with the fonts characters and their labels, whilst the supervised process is to train two different architectures of neural networks to classify these characters.

Keywords

• old documents
• ocr
• font classification
• neural networks
• romanian cyrillic