Variable Font Similarity via Design Axis Parameter Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems are unable to identify instances of variable fonts that are visually similar to a particular font, as variable fonts have a theoretically unlimited number of unique instances, making it impractical to determine visually similar instances using existing similarity score calculations based on Euclidean distances between feature vectors.
Innovation Solution
A computing device implements a similarity system that generates an instance of a variable font by modifying the value of a design axis based on attribute values of the input font, identifies additional design axes with high correlation to the similarity score, and adjusts these axes to create custom instances visually similar to the input font, reducing the need for multiple feature vector representations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional similarity score calculations based on Euclidean distances between feature vectors are used, then visual similarity between fonts can be determined, but variable fonts with theoretically unlimited instances cannot be identified as visually similar
Solution Approach 1:
The patent transforms the approach from comparing complete feature vectors to comparing specific parameters (design axis values) of variable fonts. By extracting and comparing individual design parameters such as weight, width, and slant values, the system can determine visual similarity without requiring complete feature vector representations, thus resolving the contradiction between adaptability and complexity.
2Measurement precision
If multiple feature vector representations are used to represent different font instances, then visual similarity can be calculated, but computing resource consumption increases
Solution Approach 1:
The patent extracts only the essential design parameter values from variable font instances rather than using complete feature vector representations. By taking out and comparing only the relevant design axis values (weight, width, slant), the system maintains measurement precision while significantly reducing computing resource consumption associated with processing and storing multiple full feature vectors.
3Quantity of substance
If a single variable font file is used to generate multiple instances, then font library size is reduced, but identifying visually similar instances becomes impractical
Solution Approach 1:
The patent enables detection and measurement of visual similarity by changing from comparing complete font files to comparing extracted design parameter values. By representing variable font instances through their design axis parameters rather than full font files, the system reduces the quantity of font files needed while making it practical to identify visually similar instances through parameter comparison.
Data Source
AI summary
In implementations of systems for generating instances of variable fonts, a computing device implements a similarity system to receive input data describing attribute values of glyphs of an input font. The similarity system generates a custom instance of a variable font by modifying a value of a registered design axis of the variable font based on the attribute values. A similarity score is determined that describes a visual similarity between the custom instance of the variable font and the input font. The similarity system identifies an additional design axis of the variable font based on the similarity score and generates an instance of the variable font that is visually similar to the input font by modifying a value of the additional design axis.