Real-World Font Recognition Using Deep Network and Domain Adaptation

TL;DR

This paper presents a deep learning approach with domain adaptation to improve font recognition accuracy on real-world images, bridging the gap between synthetic training data and real-world data.

Contribution

It introduces a domain adaptation technique using a Stacked Convolutional Auto-Encoder to enhance font recognition performance on real-world images.

Findings

Achieves over 80% top-5 accuracy on real-world font dataset.

Effectively reduces the domain gap between synthetic and real images.

Demonstrates the effectiveness of auto-encoder based domain adaptation in fine-grain classification.

Abstract

We address a challenging fine-grain classification problem: recognizing a font style from an image of text. In this task, it is very easy to generate lots of rendered font examples but very hard to obtain real-world labeled images. This real-to-synthetic domain gap caused poor generalization to new real data in previous methods (Chen et al. (2014)). In this paper, we refer to Convolutional Neural Networks, and use an adaptation technique based on a Stacked Convolutional Auto-Encoder that exploits unlabeled real-world images combined with synthetic data. The proposed method achieves an accuracy of higher than 80% (top-5) on a real-world dataset.