CASE_STUDIES/README.md

Simam AI Lab case studies

This collection is written for public communication, LinkedIn articles, and technical project pages. It distinguishes verified implementation evidence from research hypotheses and unfinished work.

Case studies

1. Simam3D Evidence Lab — a depth and 3D export pipeline designed to show its evidence, provenance, and uncertainty.
2. Direct Gaussian Export with DA3 — a hosted experiment exporting native Depth Anything 3 Gaussian parameters.
3. Multi-View Fusion Without Overclaiming — how duplicate, near-duplicate, synthetic-prior, and native-camera inputs are separated.
4. What's Behind the Image? — a user-guided hidden-scene hypothesis interface with explicit uncertainty boundaries.
5. SimamAnything2GS — an image/video-to-3D ingest prototype and its path toward true learned Gaussian reconstruction.
6. SimamTwin3D — a consent-first reference-capture and digital-presence research direction.

How to use these

Each case study contains:

• a short public summary;
• a LinkedIn-ready article draft;
• verified evidence and limitations;
• a suggested call to action and hashtags.

The canonical implementation and verification record are linked from the Simam3D GPU Space, including hosted verification, the paper outline, and the research backlog.

Do not describe a case study as a benchmark result unless it includes a named dataset split, a baseline, a metric, and a reproducible manifest.