Free tool · runs in your browser
Mirror Face Test
The viral test, done properly. Symmi finds your midline from 478 landmarks, then builds your left-left face and your right-right face. Underneath, the readings that explain why they look different.
Drop a photo, paste one, or choose a file
Straight-on, whole face visible, even light. JPG, PNG or WebP.
Runs in your browser. Your photo never leaves your device.
How it is measured
A measured midline, not a guess in a photo editor.
Most mirror-face pictures are made by dragging a line to roughly the middle of the nose. A few pixels off makes one composite fat and the other thin, and people read that as their real asymmetry. This tool takes the midline through the glabella and chin after removing head tilt, so both composites start from the same anatomical centre.
- Photo-left × 2
- The half on the left of the image, mirrored across the midline. On a normal (non-mirrored) photo this is your right side. Both composites are cut to the same width, so a face that sits off-centre in the frame does not make one look wider.
- Photo-right × 2
- The other half, mirrored the same way.
- Readings
- The same symmetry index and offsets as the Face Symmetry Test, so you can see which feature makes the two composites differ.
Sources
What this is based on.
The measurement and its reference ranges come from published research. Studies use different methods and populations, so compare your number with the range shown on this page rather than with other tools.
- 1.
Mita TH, Dermer M, Knight J (1977). Reversed facial images and the mere-exposure hypothesis. Journal of Personality and Social Psychology. doi:10.1037/0022-3514.35.8.597
People preferred their mirror image, while friends preferred their true image: the familiarity effect.
- 2.
Wang TT, Wessels L, Hussain G, Merten S (2017). Discriminative thresholds in facial asymmetry: a review of the literature. Aesthetic Surgery Journal. PubMed 28200081
How much left–right difference people can notice in a face.
- 3.
Kartynnik Y, Ablavatski A, Grishchenko I, Grundmann M (2019). Real-time facial surface geometry from monocular video on mobile GPUs. arXiv. arXiv:1907.06724
The face-mesh model that places the 468 surface landmarks this tool measures from.
- 4.
Ablavatski A, Vakunov A, Grishchenko I, et al. (2020). Real-time pupil tracking from monocular video for digital puppetry. arXiv. arXiv:2006.11341
The iris model that adds 10 eye landmarks (478 in total), used for pupil-based reference lines.
Our own data: how much each reading moves when the same face is photographed again is published on the methodology page. How we choose sources.
Related tools
FAQ
Why do the two faces look like different people?+
Because a mirror doubles every deviation. A 2 percent chin offset becomes a 4 percent width difference between the two composites. Everyone's mirror faces look strange; the readings tell you whether yours are actually outside the normal range.
Which side is my good side?+
Look at the readings rather than the composites. The side with the higher eye, the higher mouth corner and the fuller cheek reads as the stronger side in photos. Photographers usually shoot the other side toward the camera to balance it.
Can I fix the difference?+
Soft-tissue differences respond to targeted training: mouth corners, brow height, cheek volume. The Symmi app measures which side needs work and builds a side-aware routine. Bone differences do not change.
Is the photo stored?+
No. Everything runs in your browser and the composites are drawn on a canvas on your device.
One reading here.
Nineteen in the app.
A controlled scan, every reading against its ideal band, bone vs trainable, a plan, and a 3D model on Face ID iPhones.