The research foundation
The Face Reports Faster Than Words
This page was written for those who want to know exactly what this stands on: professionals, researchers, and anyone who simply refuses to accept promises without a basis. It includes both what the method knows and what it does not yet know.
- Duration of a micro-expression
- Up to ⅕ second
- Studies in the review
- 36
- Years of the studies
- 2021–2025
- Average accuracy in the review
- 93%
Fundamentals
What Is a Micro-Expression
A person's face is operated by dozens of muscles. Some of their movements are controlled and voluntary: a polite smile, a deliberate raised eyebrow. Others are activated before conscious control has time to intervene.
Movements of the second kind sometimes last less than a fifth of a second. They are too brief for the eye to identify consciously, and too brief for the person themselves to know they happened. A camera that captures dozens of frames per second picks them up well.
This is the starting point of the entire field: there is a layer of information in the face that does not pass through conscious decision.
The research tradition
The Facial Movement Coding System
In the 1970s, the researchers Paul Ekman and Wallace Friesen developed a systematic method for describing facial movements. Instead of describing emotions in general words, the method breaks every expression down into action units: movements of defined muscles, each of which is given a number.
The importance of this step is hard to overstate. It turned a field that depended on personal impression into one in which one can measure, repeat measurements, and compare them. This is the foundation on which research in the field stands to this day.
Our system works on the same basic logic, with today's tools: it breaks a video down into action units, measures intensity and duration, and cross-references the combinations.
From video to report
Four Stages
Breaking down the video
The video is broken down into a sequence of images. Each image is a measurement point in its own right.
Identifying action units
In each image, muscle movements are identified, along with their intensity and how long they appear.
Cross-referencing
Thousands of calculations examine which combinations recur, when they appear, and what precedes them.
Building the report
The patterns are translated into a written description of strengths, motivations, obstacles, and directions.
The main finding
What the Research Shows
A review published in the journal BMC Psychiatry examined 36 studies conducted between 2021 and 2025. The review checked how well machine-learning algorithms succeed in identifying mental states from the analysis of facial micro-expressions.
The average accuracy found across the studies stood at 93 percent.
It is important to be precise: this figure describes the performance of the algorithms in the studies reviewed, not the performance of our system. It is cited here to show that the direction is well founded, not as a promise of a result.
Why It Works Precisely Now
- CamerasEvery phone today films at a rate and quality that were once reserved for research laboratories.
- Computing powerFrame-by-frame analysis of an entire video was impossible at a reasonable cost until just a few years ago.
- Machine learningIdentifying recurring patterns across hundreds of variables at once is exactly the kind of task in which algorithms are better than the human eye.
Full disclosure
What the Method Does Not Yet Know
Anyone who presents a developing field as if it were finished is selling something. These are the limits we set for ourselves.
This is not a medical or psychological diagnosis
The assessment does not determine the existence of a disorder, illness, or medical condition, and does not replace an examination by a licensed professional.
We do not deal with lie detection
The research literature is deeply divided on this question. A field in which there is no research consensus is not one in which it is right to sell certainty.
Personality is not a final sentence
The description portrays a current structure and tendencies. It does not determine what a person can or cannot become.
The field is still developing
The foundation, the systematic coding of facial movements, is well established. Its interpretation into personal insights is an active research frontier, and we present it as such.
Frequently Asked Questions
Is micro-expression analysis proven science?
The foundation, namely the existence of tiny, involuntary facial movements and the possibility of coding them systematically, is well established and has been studied for decades. Interpreting these movements as insights into personality is a developing field, and we present it as such.
What exactly is measured?
Action units in the face: movements of defined muscles, their intensity, their duration, and the combinations between them over the course of the video. The picture is built from the cross-referencing of hundreds of such points.
Can the system detect lying?
No, and we do not deal with that. The research literature itself is deeply divided on the question of detecting lies from expressions, and anyone who promises it with confidence is selling you something.
What happens to my data?
The video is used for analysis only and is not published. Full details appear in the Privacy Policy.
A free tool as a gift
Not sure what you really want to do?
We developed a free tool that holds a deep conversation with you and helps you put into words what truly attracts you. It is sent as a gift to everyone who joins our mailing list, together with content on personality and self-knowledge.
You can unsubscribe at any time with a single click. We do not pass addresses on to anyone.