sharpetronics/QGIS
1
0
Fork 0
mirror of https://github.com/qgis/QGIS.git synced 2025-07-14 00:04:06 -04:00
Code Issues Packages Projects Releases Wiki Activity

Description of talk "QGIS Plugin AIAMAS" added

Stefan 2023-11-29 20:22:54 +01:00
parent 6cad317b39
commit 50c9e9ec0e
1 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
QOD-December-2023.md

@ -41,7 +41,8 @@ No | Name | Role | Contact
**Presenters:** Stefan Keller
**Description:**
This presentation is about a work-in-process of a QGIS plugin which uses AI to suggest map symbols (icons), including their color, for newly loaded datasets based on the dataset names. This talk aims to collect examples of layer names and feedback from the community in the early implementation phase. This plugin uses dataset names semantic search and to find matching icons. Layers are optionally reordered based on their geometric type (raster, polygon, linestring, point). Suggested colors can be adapted to a color palette for web-optimization or color-blindness. Behind the scenes, machine learning algorithms such as pre-trained large language models and clustering (sentence transformers open source from Hugging Face) are used to match layer names to colors and map icons. This is like asking ChatGPT "What is the typical color of playgrounds? Please output as RGB code in one sentence. This allows the user to quickly create visually appealing maps for web and print without tedious manual work.
This talk is about a work-in-progress QGIS plugin that uses AI to suggest map symbols (icons), including their color, for newly loaded datasets. The goal of the talk is to gather examples of layer names and feedback from the community in the early stages of implementation.
The plugin uses semantic search to match layer names with colors and map symbols/icons. Suggested colors can be adapted to a color palette for web optimization or color-blindness. Behind the scenes, machine learning algorithms are used such as pre-trained large language models and clustering (open source sentence transformers from Hugging Face). This is like asking ChatGPT "What is the color of playgrounds on a map? Please output as RGB code in a sentence". This allows the user to quickly create visually appealing maps without tedious manual work.
## Title