Detection capability
Posture & Behaviour
Posture and Behaviour analysis reads how a person holds and moves their body across a video, turning visible physical cues into one of the eight signals behind a FaceOff trust result.
What it is
The body says something the face alone does not
Posture and Behaviour is one of the eight detection capabilities that make up the FaceOff platform. It looks at visible body position and movement in a video and identifies the physical cues people commonly associate with different emotional states. The important word is visible. Nothing is inferred from anything a viewer could not also see on screen.
It also watches how a position changes rather than judging a single moment. A person can look tense in one frame and relaxed a second later, so a still image is a weak place to draw a conclusion. Happiness may show as relaxed shoulders and an upright stance. Anger may show as clenched fists and a forward lean. Sadness may show as slouching and a lowered head, surprise as lifted shoulders, fear as freezing or leaning back, and disgust as twisting away or physically withdrawing. Read across a whole clip, those changes become a signal that supports the wider result.
One of eight signals in the Trust Factor Engine. It contributes to a result and never decides one on its own. How the weighting works
How it works
From a video to an observation you can act on
- 01
Submit
Send a video the way that suits you: upload a file, connect a live feed, or pass it in from a system your team already uses.
- 02
Observe
The person's body position and movement are followed across the whole clip rather than sampled from one frame.
- 03
Interpret
Changes in stance and movement are matched to the physical cues commonly associated with each emotional state.
- 04
Combine
The observation joins the seven other signals in the Trust Factor Engine, which weighs them together rather than in isolation.
- 05
Explain
You receive a result in everyday language, with the moments in the video that shaped it, so a reviewer can check the reasoning.
What it gives you
Ordered by what changes a review first
Movement read across time
A stance is judged by how it changes through the clip, not by how it looked in the frame someone happened to pause on.
Whole clip, not one frame
Six emotional states from the body
Happiness, anger, sadness, surprise, fear and disgust each have physical cues the capability is built to recognise.
Recognised from posture alone
Works with the video you already have
Recorded files, live feeds and video arriving from your existing systems are all handled the same way. No special camera and no studio.
No new hardware needed
Always part of a wider picture
The finding is presented as one signal among eight, so nobody is asked to act on body language on its own.
One of eight signals
Explained in everyday language
Results are written for the person reviewing the case rather than for a specialist, and they point at what was seen and when.
Readable by any reviewer
Explore the eight signals
See where posture sits among the other seven
Select any signal to see what it looks at and what it adds. The Trust Factor Engine adjusts how much weight each one carries depending on what a particular video can support, so the mix is never fixed and no single signal is ever the whole answer.
You are here
Posture & Behaviour
Body position and movement, read across the whole clip.
A clear result comes from several signals agreeing. Where they disagree, the engine says so rather than picking a side quietly.
Where it is used
Anywhere a recorded person has to be understood, not just identified
Banking and insurance
Recorded claim statements and video calls carry more than words. Body language gives reviewers another reason to look closer at a claim, or to move it along with confidence.
Law enforcement
Interview and body worn footage can be reviewed for behavioural cues that help investigators decide which recordings deserve a second look.
Courts and legal
Video evidence arrives with an explanation of what was observed and when, which supports a reviewer preparing or examining a case.
Healthcare
In remote consultations, visible signs of distress or discomfort can be surfaced to a clinician who is watching a screen rather than a room.
Education
Remote assessment and supervision benefit from an extra, visible signal about how a candidate is behaving throughout a session.
Customer intelligence
Recorded conversations and video feedback can be understood in terms of how people reacted, not only what they said.
Why FaceOff
Eight signals, one result, and the reasoning left visible
No single signal decides anything
Eight capabilities are considered together, with the weight adjusted to what the footage can support. A result has to convince more than one of them.
A score with room in it
The Trust Factor Engine returns a graded trust level rather than a flat yes or no, so borderline cases look borderline instead of being forced into a box.
You can see why
Every result comes with the observations behind it in ordinary language, so your team reviews findings rather than trusting a number.
It fits how you already work
The same capability serves an uploaded file, a live feed or a system you already run, so checks happen inside normal work.
Common questions
Frequently asked questions
Posture and behaviour analysis is a way of reading visible body position and movement in a video to identify the physical cues people associate with different emotional states. In FaceOff it is one of eight detection capabilities, and it studies how a stance changes through a clip rather than judging a single frame.
No, and FaceOff does not claim it can. This capability reports what is visible in the way a person holds and moves their body. It is one input among eight, and it is designed to help a human reviewer decide where to look, not to reach a conclusion about honesty on anybody's behalf.
Six: happiness, anger, sadness, surprise, fear and disgust. Each has physical cues that tend to appear in the body, such as an upright and relaxed stance, a forward lean, a lowered head, lifted shoulders, leaning back, or turning away.
Yes. Recorded files, live feeds and video passed in from systems your teams already use are all supported. No special camera, sensor or recording setup is required.
It contributes one signal to the Trust Factor Engine, which weighs all eight together. The engine adjusts how much each signal counts based on what a given video can support, so a result reflects several capabilities agreeing rather than one speaking alone.
No. Findings are written in everyday language and point at the moments in the video that shaped them, so the person handling the case can review the reasoning without specialist training.
See it read a real video
Bring your own footage. We will run it through the platform and walk you through what each of the eight signals found, including what the body was doing and when.