/labs

Labs and evidence

Lab Notes separate method, result and limitations from customer case studies. Filter by status, domain or evidence level before making a decision.

Internal testEV-2026-004

Vision

Livestock stereo-camera sampling: internal lab note

A lab note on sampling bias when a stereo camera observes animals through one fixed scale.

Method: Compare image-derived geometry with scale observations, then check whether the sampling path covers the whole pen before interpreting accuracy.
Result: The prototype demonstrates the capture path, but the single-scale sample is too small to support a publishable weight-accuracy claim.
Limitations: A fixed single scale samples too few animals to represent the whole pen.

Reviewed: 2026-09-12Read the Lab Note

In developmentEV-2026-002

Document AI

An architecture exploration for document search that keeps the processing boundary explicit.

Method: Map payload, retrieval, logs, updates and support separately, then define the human review and citation checks needed for a pilot.
Result: The boundary can be described and reviewed, but no retrieval-recall or grounded-answer performance result is published.
Limitations: Architecture describes a deployment boundary, not a production performance guarantee.

Reviewed: 2026-09-12Read the Lab Note

Internal testEV-2026-001

Vision

An internal test of lighting geometry before model training for dark, glossy parts.

Method: Capture the same part under candidate angles, record lighting and working distance, then compare the reviewer's defect visibility before training.
Result: The geometry exercise exposed lighting as the first decision. No publishable precision, recall or throughput number has been approved.
Limitations: Lighting and camera geometry must be re-tested for each installation.

Reviewed: 2026-09-12Read the Lab Note

Lab status is part of the claim. Internal tests and in-development notes are not production guarantees.