Fine-tuning Meta SAM 3 for narrative-aware copyspace: fixing AI picture book typography
andrew_n_carr · x · 2026-09-05
Engineer andrewncarr shares a toy project that turns an episode of a kids' show into a picture book: it picks good moments, writes a short adaptation, and lays prose over the frames.
The writing side got decent quickly, but typography never did — every spread looked like a PowerPoint slide with a text block dropped onto a screenshot in whatever corner had the fewest pixels.
He frames the problem as narrative-aware copyspace: can a segmentation model propose regions that are narratively appropriate for captions or dialogue? The answer turned out to be yes. His process:
- Built a small dataset for this novel problem
- Fine-tuned Meta's SAM 3 on it
- Plugged the segmentation output into the layout pipeline
A reproducible example of adapting a foundation segmentation model to a niche layout task, transferable to other text-placement-aware image workflows.
More from coding & agent
- Vercel closed 1,500 Next.js issues in a month with a closability agent — Vjeux · 2026-09-05
- AgentMuxer launches public beta: an OpenRouter for agent capabilities with 30 providers — ycombinator · 2026-09-05
- Indie dev builds space game with hundreds of star systems using OpenAI's Astra, shares full workflow — Dimillian · 2026-09-05
- theo shares four hands-on ways to use Astra: slop audits, perf wins, and agent DX loops — lucasmeijer · 2026-09-05
- TIL: You Can Set a Token Limit on /goal in Codex — majidmanzarpour · 2026-09-05
- Meta's Muse Spark 1.3 hits #8 on Landing Page leaderboard, touted to reshape coding agent economics — yuwen_lu_ · 2026-09-05