A small number of validation cases. Selected with care.
Tattva3D is currently engaging forensic engineers, accident reconstruction teams, and insurers as pilot collaborators. The goal is mutual validation — a real case, a real measurement, a reviewable result.
Strong fit
- —Short clips — seconds to a few minutes
- —Known scene geometry, survey, or scan data available
- —A clear measurement question (e.g. speed, trajectory, position)
- —Willingness to share inputs under appropriate agreement
Possible fit
- —Partial scene geometry with measurable references
- —Calibration imagery available for the source camera
- —Cases where constraint can be added through additional measurement
Out of scope
- —Long, unconstrained, freely-moving footage with no scene anchors
- —Cases requiring inferred geometry the camera never observed
- —Requests for a single-number answer with no review of inputs
- 01
Initial conversation
A short call to understand the case, the available evidence, and the measurement that matters.
- 02
Material review
Under NDA, we review the source video, available scene data, and any calibration material.
- 03
Pilot execution
We run the constrained recovery and produce an audit manifest alongside the result.
- 04
Joint review
You — and any party you choose — can re-run the analysis from the manifest and challenge any step.
A reviewable run
An audit manifest containing inputs, parameters, intermediates, and outputs — re-executable by you or a third party.
A measurable output
A concrete result tied to the case: a speed estimate, a trajectory, or a per-frame camera pose, expressed in scene-aligned units with uncertainty.
A short technical memo
A concise written summary of assumptions, scene constraints, methods used, and the boundary of what the result supports.
Tattva3D is an independent, research-stage venture working on deterministic recovery of motion and scene-aligned measurements from monocular video. It is being built specifically for use cases where the result must withstand technical and legal review.
The current stage of the company is intentionally small and selective. Pilot collaborators are treated as design partners — their cases shape the validation roadmap, and they get direct access to the people building the system.