FULL STORY
Augmenting Descent 2e Solo Play with AI and Computer Vision
Developer TheZachMueller began experimenting with AI-driven narration and game mastering to enhance solo play of the board game Descent 2e, then progressed to stitching game boards at exact 1:1 scale using OpenCV and Claude for dungeon rendering.
2026-11-12 ~ 2026-11-12 · 2 episodes · 5 posts
Episode 1 · Developer Augments Descent 2e Solo Play with AI and Computer Vision (2026-11-12, 2 posts)
Developer TheZachMueller is documenting an experiment that combines AI narration with computer vision via an overhead document camera to recognize miniatures, delivering impressive results for solo play in the dungeon-crawling board game Descent 2e.
- Augmenting solo Descent 2e tabletop sessions with AI, CV, IR and a document camera — TheZachMueller · 2026-11-12
- Dev uses AI-tuned document camera to identify tabletop miniatures for YOLO training — TheZachMueller · 2026-11-12
Episode 2 · Dev Combines OpenCV and Claude to Rebuild 1:1 Board Game Maps and Render Dungeons (2026-11-12, 3 posts)
TheZachMueller used OpenCV and Claude to stitch 1:1 Descent 2e board maps with natural seam transitions and completed the first full dungeon render in his TTRPG engine; a follow-up RF-DETR Large detection test looked promising, though it mislabeled a monitor as a fridge.
- OpenCV edge detection plus AI blending rebuild a Descent 2e board at 1:1 scale — TheZachMueller · 2026-11-12
- Dungeon Engine's First Full Render: OpenCV Edge Detection + Claude Terrain Blending — TheZachMueller · 2026-11-12
- RF-DETR Large powers dungeon render detection — and thinks the screen is a fridge — TheZachMueller · 2026-11-12