Key terms and concepts used in emotion AI dating technology. Understanding these helps explain how Attune matches people based on genuine emotional compatibility.
A specific facial muscle movement identified in the FACS system. Each AU represents a distinct muscular action that contributes to facial expressions. For example, AU6 is the cheek raiser (orbicularis oculi), and AU12 is the lip corner puller (zygomaticus major).
Attune's EchoDepth technology analyzes 44 Action Units in real-time to understand emotional responses.
See also: FACS, Micro-expression
In the VAD model, arousal represents the intensity or activation level of an emotion. High arousal emotions include excitement and anger; low arousal emotions include calm and sadness.
One of three dimensions used to map emotional responses in Attune's matching system.
A 30-second baseline capture of your neutral facial state before emotional profiling begins. This establishes your personal reference point, allowing the system to measure deviations that represent genuine emotional responses rather than resting facial features.
Calibration ensures the system measures your emotional reactions relative to your own baseline, not against a generic standard.
In the VAD model, dominance represents the sense of control or influence associated with an emotion. High dominance emotions include confidence and contempt; low dominance emotions include fear and helplessness.
Attune's proprietary emotion AI technology that analyzes facial expressions in real-time using the device's front-facing camera. It processes 44 FACS Action Units to generate emotional profiles and compatibility scores. Named for its ability to detect the "echo" of genuine emotions beneath surface expressions.
All EchoDepth processing happens on-device. Raw video never leaves your phone.
See also: On-device Processing, FACS
The degree to which two people's emotional response patterns complement each other. Unlike physical attraction or shared interests, emotional compatibility measures how well people resonate at an involuntary, subconscious level - often a stronger predictor of relationship longevity.
Attune's core matching criterion, measured through analysis of emotional trajectories across standardized stimuli.
See also: Emotional Profile, Emotional Synchrony
A mathematical representation of your unique emotional response patterns, generated from your calibration session and stimulus reactions. Stored as anonymized VAD vectors rather than biometric data. Your profile captures how you respond emotionally - timing, intensity, duration, and recovery - not what you look like.
See also: VAD Model, Emotional Trajectory
The phenomenon where two people's emotional states begin to align during interaction. Research shows couples with higher emotional synchrony report greater relationship satisfaction. Attune measures synchrony during video dates to refine match predictions.
See also: Emotional Compatibility
The pattern of emotional change over time during a stimulus or interaction. Rather than capturing a single emotional snapshot, Attune tracks four dimensions: reaction timing (onset latency), intensity (magnitude), duration (how long it lasts), and recovery (return to baseline). This produces a "shape" that characterizes your response.
See also: Emotional Profile, Calibration
A comprehensive anatomical taxonomy of human facial expressions developed by psychologist Paul Ekman and Wallace V. Friesen in 1978. FACS breaks down facial movement into individual Action Units, allowing precise measurement of emotional expressions. It's the established standard for facial expression research across psychology, neuroscience, and human-computer interaction.
Attune uses FACS because it provides an objective, scientifically validated framework - not proprietary pseudoscience.
See also: Action Unit, Micro-expression
Technology that verifies a real person is present during the emotion capture session, not a photograph, pre-recorded video, or deepfake. Liveness detection analyzes micro-expressions, natural movement patterns, and response timing to ensure authenticity.
This is how Attune guarantees zero fake profiles - bots and catfish cannot fake involuntary micro-expressions.
See also: Micro-expression
A brief, involuntary facial expression that occurs in under 200 milliseconds. Unlike deliberate expressions, micro-expressions are extremely difficult to consciously control or fake. They reveal genuine emotional responses before the conscious mind can intervene.
EchoDepth's sub-200ms detection latency allows it to capture these fleeting but authentic emotional signals.
See also: Action Unit, Liveness Detection
A privacy architecture where facial video analysis happens entirely on your smartphone. Raw biometric video never leaves your device or gets transmitted to servers. Only the resulting mathematical emotional profile (anonymized VAD vectors) is stored.
This design ensures your facial data remains private while still enabling emotional compatibility matching.
See also: EchoDepth, Emotional Profile
The most widely used dimensional model of emotion in affective computing research. Rather than categorizing emotions into discrete labels (happy, sad, angry), VAD maps emotional states across three continuous dimensions: valence (positive to negative), arousal (calm to excited), and dominance (submissive to dominant).
This model allows Attune to capture the full spectrum of emotional response rather than reducing feelings to simple categories.
In the VAD model, valence represents the positive-to-negative quality of an emotion. Positive valence includes joy, interest, and affection; negative valence includes sadness, disgust, and fear.
Want to understand the science in more depth? Read our full methodology or explore research insights.