How
It Started.
In 2021, we were working with a database of thousands of people who had taken personality tests and shared their portraits. Something kept catching our attention. People with the same personality type looked similar. Not identical — but noticeably, measurably similar. In some types, the visual alignment reached 78%.
We started asking — What if personality really does leave visible traces?
Over the next four years, we expanded the database to 380,000 portraits. We developed computer vision models and facial analysis algorithms, identifying 126 distinct parameters to track muscle patterns, expressions, and energy signatures. What we found surprised us.
Most people share common visual traits — definable and measurable — that align with their behavioural patterns. The way they express emotion. Where their gaze naturally goes. The muscle patterns developed from thinking the same way.
In the last few decades, famous clinical psychologist Paul Ekman proved that people's emotions are expressed on their faces. Our research that people's cognitive processes and behavioural patterns are reflected in people's faces.
How
is it going.
The research is still in progress. But the patterns are already clear enough to see and use. The database is substantial, but it's not complete. We are working on expanding it, but there are real limitations.
Age.
Our research focuses on adults between roughly 25 and 65. Before 25, personality is still forming. After 65, life experience and aging can mask underlying patterns. We don't have enough data to make confident claims outside this range.
Race / Culture.
Our database is Western-based or European since it has 87% of. We don't yet have enough visual data from East Asian, South Asian, African, Latin American, or Middle Eastern cultures. Facial expression norms vary across cultures — eye contact, emotional display, and how intensity is read. The underlying cognitive patterns may be universal, but how they look likely varies.
Cosmetic change.
Extensive cosmetic interventions like Botox, fillers, or facelifts can alter facial expressions, muscle patterns, and resting features. These changes can mask or distort natural indicators, making VPTI typing unreliable for individuals with substantial procedures. In such cases, we recommend using behavioral observation or traditional typing methods instead.
What
is will be.
We're sharing what we've found so far — where our data is strong and where we're still learning. If you'd like to help expand the database, we'd appreciate it. (how?) The more diverse our data becomes, the more accurate VPTI will be.
We're sharing what we've found so far — where our data is strong and where we're still learning. If you'd like to help expand the database, we'd appreciate it. (how?) The more diverse our data becomes, the more accurate VPTI will be.