EV
Camera event ingest
LiveAttach detected or manually marked events to the match.
DCS combines official scoring, ordinary cameras, derived analytics, model estimates and authorised human review—without allowing AI to silently become the official record.
The current layer captures, surfaces and reviews match evidence honestly.
Attach detected or manually marked events to the match.
Confidence-scored clips for boundaries, wickets and catches.
Human confirms or rejects consequential events.
Signed bundle of reviewed match decisions.
Every output carries source, confidence, angle and limitation metadata.
2D reconstructed match state and event timeline.
Scoring zones, matchups and expected next-over context.
Ordinary-camera movement observations—not diagnosis.
Starting position, reaction, route and throw evidence.
These modules require calibrated multi-camera data, independent validation and governance before production use.
Multi-camera release, bounce and trajectory reconstruction.
Fixed crease camera with ABSTAIN as a valid output.
Synchronised audio/video contact estimate.
Recommendation only; authorised human remains final.
Candidate → threshold passed → independent review → founder approval → limited release → production.
Start with scoring and smart streaming, then expand to controlled camera intelligence.