Font Creation Apparatus Style Transfer Across Scripts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing font creation systems cannot generate a complete font set for languages like Hiragana, Katakana, or other scripts with a desired style, as they are limited in transforming standard fonts into unique handwriting styles or imitating specific alphabets.
Innovation Solution
A font creation apparatus that acquires character images, extracts feature amounts, estimates transformation parameters between the extracted features and a reference font, and generates a complete font set by transforming a reference font set using these parameters, enabling the creation of a complete font set for any language with a desired style.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a handwriting-style character font generation system transforms a standard character font into a feature amount representing user individuality by extracting displacement of corresponding points, then a unique handwriting-style character font can be generated, but it is not possible to generate a handwriting-style Katakana character font by using the handwritten Hiragana of the user
Solution Approach 1:
The patent introduces a complete reference font set as an intermediary between the user's handwritten characters and the target complete font set. The transformation process uses this reference font set to bridge the gap, allowing style transfer across different languages and character sets without requiring direct correspondence between input and output character types.
Solution Approach 2:
The patent segments the font creation process into distinct modules: feature amount extraction from user input, transformation parameter estimation between features, and font generation from transformed features. This segmentation allows the system to handle different languages and character sets independently while maintaining a unified transformation framework.
2Quantity of substance
If a GAN generates 26 alphabet characters having a style similar to small-number alphabets by using generative adversarial networks, then a styled font can be generated from few samples, but it is not possible to generate a complete font set for Hiragana, Katakana, or another language
Solution Approach 1:
The patent creates a universal transformation framework that can generate complete font sets for multiple languages including Hiragana, Katakana, and other scripts. The system uses a complete reference font set that encompasses various language character sets, allowing the same transformation process to be applied universally across different languages and writing systems.
Solution Approach 2:
The patent prepares a complete reference font set in advance that includes all necessary character types for different languages. This preliminary preparation enables the system to generate complete font sets for any language by simply transforming the reference set using estimated transformation parameters, without requiring language-specific processing or additional training data.
3Measurement precision
If feature amount extraction extracts displacement of corresponding points between handwritten character and standard character, then individuality of user can be captured, but corresponding points must exist between different character types which limits cross-language generation
Solution Approach 1:
The patent uses a complete reference font set as an intermediary that provides corresponding points for all character types across different languages. Instead of directly matching corresponding points between user's handwritten Hiragana and target Katakana characters, the system extracts features from the reference font set which serves as a universal mediator, enabling accurate feature extraction and transformation across any language boundary.
Data Source
AI summary
There are provided a font creation apparatus, a font creation method, and a font creation program capable of generating, from small-number character images having a desired-to-be-imitated style, a complete font set for any language having the same style as the character images. A feature amount extraction unit (40) receives a character image (32) of a first font having a desired-to-be-imitated style and extracts a first feature amount of the first font of the character image (32). An estimation unit (42) estimates a transformation parameter between the extracted first feature amount and a second feature amount of a reference second font (34). A feature amount generation unit (44) generates a fourth feature amount of a second font set to be created by transforming a third feature amount of a complete reference font set (36) based on the estimated transformation parameter. A font generation unit (46) generates a complete second font set by converting the generated fourth feature amount of the second font set into an image.