Convolutional Neural Networks for Font Classification

TL;DR

This paper introduces a CNN-based framework for font classification that achieves state-of-the-art accuracy on Arabic fonts and medieval Latin manuscripts, with analysis of learned features and improved robustness through data augmentation.

Contribution

The paper presents a simple CNN-based method for font classification that outperforms previous approaches and introduces a novel data augmentation technique for better robustness.

Findings

Achieved 98.8% line accuracy on Arabic fonts

Achieved 86.6% page accuracy on Latin manuscripts

Proposed data augmentation improves robustness to text darkness

Abstract

Classifying pages or text lines into font categories aids transcription because single font Optical Character Recognition (OCR) is generally more accurate than omni-font OCR. We present a simple framework based on Convolutional Neural Networks (CNNs), where a CNN is trained to classify small patches of text into predefined font classes. To classify page or line images, we average the CNN predictions over densely extracted patches. We show that this method achieves state-of-the-art performance on a challenging dataset of 40 Arabic computer fonts with 98.8\% line level accuracy. This same method also achieves the highest reported accuracy of 86.6% in predicting paleographic scribal script classes at the page level on medieval Latin manuscripts. Finally, we analyze what features are learned by the CNN on Latin manuscripts and find evidence that the CNN is learning both the defining morphological differences between scribal script classes as well as overfitting to class-correlated nuisance factors. We propose a novel form of data augmentation that improves robustness to text darkness, further increasing classification performance.