Research and development of generative neural networks for type design
Academy of Fine
Arts and Design in Bratislava
Department of Visual
Communication
Studio Typo
Filip Paldia
Supervisor: Pavol Bálik
Consultant: Zuzana Panák
Husárová
Bratislava 2025
Contents
- Title Page
- Abstract
- 1. Introduction
- 1.1. Context: Complaining about Font Production Limitations
- 1.2. Objectives: The Purpose of the Project
- 1.3. Contributions: Place for Praise
- 1.4. Outputs: Tangible Artefacts of the Project
- 1.5. Overview: What is in it and what is not
- 2. Preliminaries: The Inception of the Research
- 2.1. Exploration: Collection of Bibliographic Materials
- 2.2. Systematisation: The Taxonomy of AI Font Generation Projects
- 2.3. Compilation: The Database of AI font Generation Projects
- 3. Surveys: The State of the Domain
- 3.1. Applications: The Use of Generative Models in Type Design
- 3.1.1. Font Blending
- 3.1.2. Font Completion
- 3.1.3. Cross-modal Generation
- 3.2. Fonts Datasets
- 3.2.1. Bitmap Image Font Datasets
- 3.2.2. Real Font Datasets
- 3.2.3. Observations
- 3.3. Representations: Methods of Encoding Font Data into Numerical Values
- 3.3.1. Brief Journey Into Numerical Representation
- 3.3.2. Font Data Nature
- 3.3.3. Spatial Representation
- 3.3.4. Sequential Representation
- 3.3.5. Multimodal Representation
- 3.4. Discussion: Interlude to the Practicalities
- 4. Projects: Practical Applications
- 4.1. Proposal: Down of Regularised Font Datasets
- 4.2.1. Methods
- 4.2. Framework: Round Trip Through Dataset Factory
- 4.2.2. Tools
- 4.3. Presentation: Visit the Product Store
- 4.4. Demonstration: Use Case of the Dataset
- 4.4.1. Method
- 4.4.2. Results
- 4.5. Evaluation
- 4.5.1. Heuristics
- 4.5.2. Individual Evaluation
- 4.5.3. Comparative Evaluation
- 5. Conclusions
- 6.1. Definitions
- 6.2. Abbreviations
