Exploring HENet: Forcing a Network to Think More for Font Recognition: A Brief Overview

Author(s): Vincent Liu

Originally published on Towards AI.

An end-to-end font recognition network
Source: Image by Anatoly777 from Pixabay

How to identify the font of the text in an image? For one comes from the image processing background may say: Template Matching. Others who don't mind moving one step further may consider classification networks.

Identifying the font could be straightforward when the text is aligned and clear. However, it becomes challenging when the image quality varies and the text is skewed or rotated. Moreover, there are thousands of available fonts, and many look extremely similar with only subtle differences.

How about existing computer vision solutions? There are many font identifier online services where users can upload a text image, manually crop the letters, and modify thresholds of brightness and contrast to make the text stand out. The services usually return a long list of matched fonts sorted by confidence score. Instead of returning the most confident font, the returned list of potentially matched fonts somewhat answers the question.

In this article, we would like to share the note of the paper: HENet: Forcing a Network to Think More for Font Recognition¹. This is one of the innovative yet comprehensible works in font recognition published in recent years. Let’s delve into the idea of this paper in the… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI