Analyzing Font Style Usage and Contextual Factors in Real Images

Naoya Yasukochi,Hideaki Hayashi,Daichi Haraguchi,Seiichi Uchida

2023-06-21

Abstract:There are various font styles in the world. Different styles give different impressions and readability. This paper analyzes the relationship between font styles and contextual factors that might affect font style selection with large-scale datasets. For example, we will analyze the relationship between font style and its surrounding object (such as ``bus'') by using about 800,000 words in the Open Images dataset. We also use a book cover dataset to analyze the relationship between font styles with book genres. Moreover, the meaning of the word is assumed as another contextual factor. For these numeric analyses, we utilize our own font-style feature extraction model and word2vec. As a result of co-occurrence-based relationship analysis, we found several instances of specific font styles being used for specific contextual factors.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper attempts to address the issue of understanding the relationship between font styles and their contextual factors. Specifically, the authors aim to reveal patterns of font selection in different scenarios and the reasons behind them through the analysis of large-scale datasets. The research mainly focuses on the following aspects:

1. **The relationship between font styles and contextual factors**: By analyzing the use of fonts in actual images, the study explores the relationship between font styles and surrounding objects (such as the word "TOWN" on a bus) as well as the meaning of the text.

2. **Scene images and book cover analysis**: The study examines two different application scenarios—scene images (such as street signs) and book cover images. These two scenarios provide different contextual factors, helping to comprehensively understand the logic behind font selection.

3. **Co-occurrence analysis**: Through co-occurrence analysis using large-scale datasets (such as the Open Images dataset), the study discovers patterns of specific font styles appearing in specific contextual factors. For example, handwritten fonts often co-occur with whiteboards; children's book covers tend to use specific types of fonts, etc.

The main contribution of the paper is the first attempt to systematically analyze the relationship between font styles and multiple contextual factors (objects, types, and semantics) using large-scale datasets. It also proposes a method for font feature extraction based on convolutional neural networks and word embedding techniques to quantify these relationships. Through these analyses, valuable empirical knowledge can be provided for future design work, and it can help understand human behavior patterns in font selection.