AI face mapping for permanent makeup

AI face mapping for faster, more consistent brow measurement in permanentmakeup

Tech stack: Python / OpenCV / Expo / React Native / FastAPI / MediaPipe / PostgreSQL

Background

Before any permanent makeup (PMU) treatment, such as microblading, brow shaping ortinting, the eyebrows have to be measured. This step, called brow mapping, decideswhere each brow starts, peaks and ends.

PMU artists usually do this by hand. They find the glabella, the center point between thebrows, mark the start, arch and tail of each brow with a ruler and pencil, and then checkboth sides for symmetry. On average, this takes around 30 minutes per client, before theactual work begins.

A mobile app that turns this step into a single photo, taking the measurements andproducing a brow stencil fitted to each client's face, gives that time back to the artist.

Challenge

Manual brow mapping is slow because it is precise work on a small area. The app had to meet three conditions:

One unmarked reference point. Every measurement starts from the glabella, andnothing on the face marks it. A few millimeters off and both brows end up asymmetric.
Every kind of brow. Full, thin, over-plucked or missing entirely, the mapping has towork on all of them without manual correction.
The artist's pace. The photo is taken on a phone, and the stencil has to be ready beforethe client sits in the chair.

Solution

ZegaSoftware designed a mobile app that reads a single photo of the client's face, locatesthe glabella and maps where each brow should start, peak and end. These are the samepoints the artist used to mark by hand.

Brow points from the structure of the face

The app uses AI face landmark detection, a computer vision technique that places thebrow points based on the geometry of the whole face rather than on the brow hair.Because of this, the mapping does not depend on how much brow the client has, and itworks on faces with no eyebrows at all. The mapping rules follow the artist's own method,defined together with her.

From brow map to stencil

From those points, the app generates a symmetric eyebrow stencil proportioned to theclient's face. The artist can print it on a thermal printer over Bluetooth or export it as aPDF. A live camera overlay lets her check the placement on the client before she starts.

Hosted in the EU

Client photos and data are stored and processed on infrastructure based in the EU.

Result

What it changes for the artist

Replace around 30 minutes of manual brow mapping with a single photo.
Map every face, including clients with thin, sparse or no eyebrows, without manualcorrection.
Produce consistent stencils, generated from the same points every time.

Conclusion

This project shows where AI makes a difference for a small business: a precise task thatrepeats with every client. The artist keeps her craft, and brow mapping that used to takehalf an hour starts from a single photo.

Case studies