How It Works
From Face
to Decision.
In Seconds.
Five steps. Forty-four Action Units. One platform that tells coaching staff, welfare officers, and medical teams exactly what they need to know — at the moment they need it.
01
Face · Voice · Text
Multimodal Input Capture
EchoDepth Sport captures facial video as its primary input — passively, during training sessions, warm-ups, half-time talks, and match-day environments. Voice and text inputs can also be integrated. No wearables, no disruption to session flow.
02
FACS · Real-time · Sub-second latency
44 Action Unit Detection
The FACS engine analyses each frame of facial video against the 44 Action Units it tracks — from brow lowering (AU 4) to lip corner pulling (AU 12) to blink rate (AU 41/42). Each AU activation is scored with sub-second latency.
03
Valence · Arousal · Dominance
VAD Emotional Modelling
Action Unit patterns are mapped to the three-dimensional VAD emotional model — Valence (positive/negative), Arousal (activation level), and Dominance (control/confidence). This converts raw facial data into interpretable emotional state.
04
Individual delta · Not population average
Individual Baseline Comparison
Every player's VAD scores are compared against their own individual baseline — calibrated over their first 3–4 weeks on the platform. Deviations from a player's personal norm are flagged, not deviations from a population average. Every player is different.
05
Coach · Welfare · Medical
Actionable Insight Delivery
Plain-language insights are delivered to the right staff member at the right moment. Half-time briefs for coaches. Welfare alerts for welfare officers. Impact assessment packs for medical staff. No data science degree required.
44
Action Units tracked
Of the 46 defined in the FACS standard
3D
Emotional model
Valence, Arousal, Dominance — not just positive/negative
∞
Individual baselines
Every player's unique norm — not population average
How It Works FAQs
Does EchoDepth Sport require specialist hardware?
No. EchoDepth integrates with standard video capture equipment available in most professional sport environments. Specific requirements are confirmed during the scoping process.
How accurate is facial Action Unit detection in sport environments?
EchoDepth Sport is calibrated for real-world sport conditions including lighting variation, partial occlusion, and movement. Accuracy is validated against the player's own individual baseline, which accounts for their specific expression patterns.
What is the latency from capture to alert?
Sub-second for Action Unit scoring. Alert delivery to staff dashboards is near real-time, typically within a few seconds of threshold breach.