[Paper Review] Serif or Sans: Visual Font Analytics on Book Covers and Online Advertisements

Yuto Shinahara, Takuro Karamatsu+3 more|arXiv (Cornell University)|Jun 24, 2019

Handwritten Text Recognition Techniques24 references4 citations

TL;DR

This paper proposes an automated visual font analytics framework that extracts font style, color, and genre from 207,572 book covers and 30,000 online advertisements using scene text detection, style classification, and clustering. It reveals statistically significant trends: sans-serif dominates in business and financial content, serif is preferred in fashion and lifestyle genres, and color usage correlates strongly with target audience and emotional tone, demonstrating systematic design choices in typography for genre-specific impression and atmosphere.

ABSTRACT

In this paper, we conduct a large-scale study of font statistics in book covers and online advertisements. Through the statistical study, we try to understand how graphic designers relate fonts and content genres and identify the relationship between font styles, colors, and genres. We propose an automatic approach to extract font information from graphic designs by applying a sequence of character detection, style classification, and clustering techniques to the graphic designs. The extracted font information is accumulated together with genre information, such as romance or business, for further trend analysis. Through our unique empirical study, we show that the collected font statistics reveal interesting trends in terms of how typographic design represents the impression and the atmosphere of the content genres.

Motivation & Objective

• To uncover hidden statistical trends in font usage across different content genres in visual designs such as book covers and online advertisements.
• To develop an automated pipeline for extracting font style, color, and genre information from real-world graphic designs at scale.
• To provide quantitative evidence for how typographic choices (font style and color) influence perceived atmosphere and genre representation.
• To identify systematic design principles that typographers implicitly follow, such as font preferences for specific genres or emotional effects.
• To enable data-driven applications in font recommendation, generation, and typography evaluation by revealing empirical patterns in real-world design data.

Proposed method

• Apply scene text detection to isolate titles from book cover images, followed by character recognition and font style classification into nine typographic categories.
• Use multi-dimensional scaling and clustering techniques to analyze relationships between font usage patterns and content genres.
• Extract font color by applying color quantization to segmented text regions, using eight representative color clusters.
• For online advertisements, leverage vector file metadata to obtain precise font names and colors, bypassing detection and classification for higher accuracy.
• Aggregate font style and color frequencies per product and genre, then average across products to compute genre-level statistics.
• Apply statistical analysis and visualization to reveal correlations between font characteristics (style, color) and business or content genres.

Experimental results

Research questions

• RQ1 How do font styles (e.g., serif, sans-serif, calligraphy) vary across different content genres in book covers and online advertisements?
• RQ2 What is the relationship between font color and genre, and how does it contribute to the perceived atmosphere of the design?
• RQ3 Are there systematic patterns in font usage that reflect genre-specific design intentions or psychological impressions?
• RQ4 How do media formats (e.g., book covers vs. online ads) influence font and color preferences in visual design?
• RQ5 To what extent do typographers follow implicit rules in font selection that align with genre and target audience?

Key findings

• Sans-serif fonts are the dominant style across all online advertisement genres, with over 90% usage in 'Business and Finance' and 'Personal Finance' categories.
• Serif fonts are significantly preferred in 'Style & Fashion' and 'Healthy Living' genres, with the lowest sans-serif frequency in 'Style & Fashion' (only 15% of texts).
• Calligraphy (brush-style) fonts are frequently used in 'Video Gaming' ads, particularly for titles related to historical or narrative-driven games.
• Warm and bright colors—especially red and brown—are dominant in advertisements, with red used for emphasis in 'Personal Finance' and 'Content Channel', and brown in 'Style & Fashion'.
• Blue colors are rarely used except for dark blue in 'Healthy Living', suggesting a strategic use of color to convey trust or calmness.
• Pink is used in 'Style & Fashion' not as a substitute for red but to create a feminine, atmospheric tone, indicating that hue similarity does not imply functional equivalence in design.