App · updated 2026-07-15
In progressMineral ID AI
Visual mineral identification
Computer visionMaterials science
Overview
Image-based mineral identification tool, spun out of instrumentation and materials-characterization work in the lab.
Architecture
- Image-based classification pipeline for visual mineral identification
- Spun out of lab instrumentation and materials-characterization work
- A full internal spec (MINERAL_ID_AI_SPEC.md) already exists — this page is the public summary and should be kept in sync with it, not treated as the source of truth
Milestones
- 01Bring this public summary in line with the internal spec doc
- 02Build the initial classifier on a labeled mineral image set
- 03Wire into the Tardigrade Innovation demo console alongside Tardigrade Shield
Open questions
- Which mineral classes are in scope for v1
- Data source and labeling strategy for training images
Early access
Want to try mineral identification from a photo? Leave your email for early access once the classifier is trained.