# 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](01-simam3d-evidence-lab.md) — a depth and 3D export
   pipeline designed to show its evidence, provenance, and uncertainty.
2. [Direct Gaussian Export with DA3](02-da3-direct-gaussian-export.md) — a
   hosted experiment exporting native Depth Anything 3 Gaussian parameters.
3. [Multi-View Fusion Without Overclaiming](03-multiview-fusion-evidence.md) —
   how duplicate, near-duplicate, synthetic-prior, and native-camera inputs are
   separated.
4. [What's Behind the Image?](04-whats-behind-the-image.md) — a user-guided
   hidden-scene hypothesis interface with explicit uncertainty boundaries.
5. [SimamAnything2GS](05-anything2gs-ingest.md) — an image/video-to-3D ingest
   prototype and its path toward true learned Gaussian reconstruction.
6. [SimamTwin3D](06-samamtwin3d-consent-first-capture.md) — 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](https://huggingface.co/spaces/junaid-simamdigital/Simam3D-GPU),
including [hosted verification](../HOSTED_VERIFICATION.md), the
[paper outline](../PAPER_OUTLINE.md), and the
[research backlog](../RESEARCH_BACKLOG.md).

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.