2.2. Systematisation: The Taxonomy of AI Font Generation Projects

The exploration component presents challenges analogous to navigating a sprawling, disordered archive. Raw codes and memos necessitate extensive organisation, while an overwhelming number of undefined terminologies complicate comprehension. Frustration arises from inconsistent term usage across the literature, underscoring the imperative for a systematic structuralisation of a taxonomy and definition of terms. Such a structured methodology is essential to streamline information and facilitate coherent communication within the research framework.

The establishment of standardised definitions draws upon two complementary literary sources. The seminal textbook Deep Learning (Goodfellow, Ian, Bengio, Yoshua, and Courville, Aaron 2016) provides comprehensive foundational knowledge despite its limitations in vector graphics coverage and pre-2016 scope. Survey papers (Achraf Oussidi and Azeddine Elhassouny 2018; Lin et al. 2021; Sarker 2021; Xiaojie Guo and Liang Zhao 2022; Yang 2024; Zou et al. 2024) complement this foundation by addressing recent technological developments within the specific context of font generation and vector graphics processing.

The taxonomy is developed around understanding 1.) The pipeline of converting font files (e.g. .otf, .ttf) into generative models, encompassing input (standard font files), transformation (data processing and formatting), and output (trained models) 2.) Deployment of emerged models and their application in type design workflow incorporation.

The analysis yielded 18 distinct topics with associated definitions and terminology. Mapping was implemented to consolidate terminology across the literature, much like an art curator creating a concordance between different names for the same painting across multiple museum catalogues. Example aliases were documented to connect variant terms to their standardised definitions.

The mapping focused particularly on three parent classifications that exhibited the most terminological variation:

  1. Task
  2. Encoding/Decoding Modality
  3. Architecture.
Table presents the taxonomy of AI font generation projects in 18 observed research topics
Achraf Oussidi, and Azeddine Elhassouny. 2018. “Deep Generative Models: Survey.” In 2018 International Conference on Intelligent Systems and Computer Vision (ISCV), 1–8. https://doi.org/10.1109/ISACV.2018.8354080.
Goodfellow, Ian, Bengio, Yoshua, and Courville, Aaron. 2016. Deep Learning. MIT Press. https://www.deeplearningbook.org/.
Lin, Tianyang, Yuxin Wang, Xiangyang Liu, and Xipeng Qiu. 2021. “A Survey of Transformers.” June 15, 2021. https://doi.org/10.48550/arXiv.2106.04554.
Sarker, Iqbal H. 2021. “Deep Learning: A Comprehensive Overview on Techniques, Taxonomy, Applications and Research Directions.” SN Computer Science 2 (6): 420. https://doi.org/10.1007/s42979-021-00815-1.
Xiaojie Guo, and Liang Zhao. 2022. “A Systematic Survey on Deep Generative Models for Graph Generation.” October 4, 2022. http://arxiv.org/abs/2007.06686.
Yang, Chen. 2024. “The Investigation of Deep Learning Models Utilized in Vector Graphics Manipulation.” Highlights in Science, Engineering and Technology 85 (March):1115–21. https://doi.org/10.54097/t057w194.
Zou, Bocheng, Mu Cai, Jianrui Zhang, and Yong Jae Lee. 2024. “VGBench: Evaluating Large Language Models on Vector Graphics Understanding and Generation.” In. https://arxiv.org/pdf/2407.10972.

License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License .

Creative Commons License

Citation

If this work is useful for your research, please cite it as:

@phdthesis{paldia2025generative,
  title={Research and development of generative neural networks for type design},
  author={Paldia, Filip},
  year={2025},
  school={Academy of Fine Arts and Design in Bratislava},
  address={Bratislava, Slovakia},
  type={Doctoral thesis},
  url={https://lttrface.com/doctoral-thesis/},
  note={Department of Visual Communication, Studio Typo}
}