Preprint+Article: Generative text-to-image diffusion for automated map production based on geosocial media data
2024-02-13 - The final article is online:
Dunkel, A., Burghardt, D. & Gugulica, M. Generative Text-to-Image Diffusion for Automated Map Production Based on Geosocial Media Data. KN J. Cartogr. Geogr. Inf. (2024). https://doi.org/10.1007/s42489-024-00159-9
Link to preprint: https://www.researchsquare.com/article/rs-3503977/v1
After my tests with Stable Diffusion on the TUD HPC cluster, we were confident that we could implement a fully automated data processing pipeline for map production. This preprint describes the individual components. There are many possible improvements at each stage, but we wanted to demonstrate a single integrated process from data processing to map visualization before moving on to optimizations.
Preprint:
Dunkel, A., Burghardt, D., Gugulica, M. (2023). Generative text-to-image diffusion for automated map production based on geosocial media data, 03 November 2023, PREPRINT (Version 1), Springer. DOI: 10.21203/rs.3.rs-3503977/v1
The Supplementary Materials with data, code, Jupyter notebooks and additional graphics are available in a separate data repository:
Dunkel, A., Burghardt, D., Gugulica, M. (2023). Supplementary materials for the publication “Generative text-to-image diffusion for automated map production based on geosocial media data.” DOI: 10.25532/OPARA-253
The most recent updates of these files can be found in the Gitlab repository.
Links to HTML-converted Notebooks:
- Notebook 1: Data preparation, spatial tag clustering (Tagmaps)
- Notebook 2: Generative AI parameters & tests
- Notebook 3: Map generation