The Science.
Documented.
Technical methodology documentation for EchoDepth Sport's FACS-based emotion detection system. Full transparency on how we analyse facial Action Units, derive VAD scores, and produce actionable emotional state metrics.
By Jonathan Prescott, Founder & CEO — MBA (Bayes Business School), B.Eng Computer Systems Engineering. Last updated 13 August 2026.
Why methodology transparency matters
Most emotion-AI vendors treat their detection model as a proprietary black box. EchoDepth Sport takes the opposite approach: every methodology document on this page describes the exact pipeline — from raw video ingestion through to the final emotional state score — that produces the numbers your welfare and coaching staff act on.
This matters for three reasons. First, auditability: a welfare officer or club doctor making a duty-of-care decision needs to understand what evidence underpins an alert, not just trust a score. Second, defensibility: methodology built on the peer-reviewed FACS standard (Ekman & Friesen, 1978) and VAD emotional modelling (Russell, 1980) can be independently scrutinised — see our full research and evidence base. Third, calibration: our Action Unit detection model has been calibrated across 14 cultural cohorts, and that calibration process is documented rather than asserted.
The methodology documents below cover specific analysis pipelines in full technical depth. For the underlying scientific foundation, see The Science.
Technical Deep Dives Available
If you require additional technical documentation — validation studies, cultural calibration data, or API specifications — contact our research team.
Contact Research Team