How do I know which treatment is right for me?

Book a complimentary consultation with our clinicians. We assess your skin and goals before recommending anything, and there is no pressure to proceed.

Are these treatments safe?

NexGen only offers evidence-based treatments administered by qualified, insured clinicians. We follow UK CAP/ASA and MHRA guidance and never advertise prescription-only medicines.

Do you offer aftercare?

Yes. Every treatment includes clear aftercare instructions and a review appointment where clinically appropriate.

One Photo vs 3D Scans: What Facial Symmetry Analysis Misses

An online facial symmetry analysis gives you a photographic estimate of left-right balance, useful as a curiosity or a first look, but it is not a clinical measurement. Automated landmarking carries a measurable error, as a systematic review of automated landmark accuracy shows, and clinical 3D assessment can provide a more detailed view. Sudden asymmetry follows NHS FAST stroke guidance, and our clinic can explain the rest.

TL;DR:

A 2023 review found automated landmarking differed from manual placement by 0.67–4.73 mm on average, and scores vary across tools because definitions differ.

Scores vary by tool; among 270 adolescents, a 3D study found averages of 53.49% for males and 58.50% for females, not universal targets.

A total score can conceal a conspicuous difference in one region, so inspect individual feature readings and consider whether expression, shadows, or hair shifted landmarks.

Before uploading, check whether the tool processes your photo in your browser or stores it on a server, because facial images are personal data.

Call 999 immediately for sudden facial drooping, arm weakness, or speech trouble; do not wait for symptoms to pass or use a photo to check.

How is facial symmetry measured with landmarks, midlines and 3D scans?
What do tools score on the eyes, brows, nose, mouth and jaw?
What does a symmetry score actually mean?
Where can consumer symmetry tools go wrong?
How do you get a cleaner result and keep your photos private?
When is facial asymmetry a medical emergency?
Why do clinicians use 3D imaging and holistic assessment?
How does symmetry analysis fit with other facial harmony assessments?
What are common asymmetry patterns and how are they usually corrected?
How do we manage expectations about symmetry improvements?
Our view: symmetry is one input, not a verdict
FAQ
Sources

How is facial symmetry measured with landmarks, midlines and 3D scans?

Every symmetry tool starts from the same idea: find fixed points on the face, draw a vertical midline, and compare the left side with the right. What differs is how those points are found and how much you can trust them.

Consumer tools work from a flat photograph. Software detects landmarks automatically, tracing the eyes, brows, nose, lips and jawline. It then builds an estimated midline, pairs each left-hand point with its right-hand twin, and measures the residual distance between them once the face has been mirrored. A small residual reads as balance, a large one as difference.

Clinical methods work in three dimensions. Laser surface scanning and 3D imaging capture the shape of the face itself, which is why laser surface scanning has been used in research to quantify symmetry in adolescents. Clinicians and researchers then align the scan to a reference plane and compare the two halves across the whole surface, not just at a handful of points.

The key differences in practice:

Landmark source: consumer tools infer points from pixels, while clinical protocols use anatomical definitions based on curvature and curves, which reduces reliance on head orientation.
Midline construction: a 2D tool estimates the midline from facial points, so a slight head turn shifts it. Clinical 3D protocols align on stable structures, for example the basilar surface and iterative closest point methods, to estimate the symmetry plane.
What is captured: a photograph records a single viewing angle, whereas a 3D scan records depth, contour and volume.
Reproducibility: repeat a clinical scan under a standard protocol and you should get a comparable result. Repeat a selfie in different light and you may not.

None of this makes online tools useless. It means they answer a narrower question: how balanced does this particular photograph look to this particular algorithm?

What do tools score on the eyes, brows, nose, mouth and jaw?

Most online analysers break the face into regions and report a score for each. Knowing what is measured helps you read the result sensibly.

Eyes: tools compare eye height, the tilt of the line between inner and outer corners (canthal tilt), and the distance from each eye to the midline.
Brows: the usual measures are brow height above the eye, arch position and overall slope on each side.
Nose: the bridge is checked for deviation from the midline, and the nostrils and nasal tip for lateral shift. Small shifts matter more than you might expect, since as little as 2 mm in nasal proportions can change how similar two faces are rated.
Mouth: software looks at the height of the lip corners, the position of the cupid's bow relative to the midline, and how the upper and lower lips sit against each other.
Jaw and outline: the jawline contour, the angle of each jaw, and the position of the chin point are compared across the midline.

A single weak region can matter more than the overall number suggests. An aggregate score averages everything, so one noticeable difference around the mouth can sit comfortably inside a respectable total. If a particular feature draws your eye, the feature-level reading is more informative than the headline figure.

Many differences in these readings have nothing to do with your anatomy. A slight smile lifts one lip corner. A raised eyebrow changes brow height for that frame. Light from a window casts a shadow that shifts the apparent edge of the jaw. Hair across one cheek can move a landmark entirely. Before reading anything into a feature score, ask whether the photograph itself could explain it.

What does a symmetry score actually mean?

A symmetry score is a single number built from many small comparisons, and it needs context before it means anything. Percentages vary between tools because each one defines and weights symmetry differently, so a score from one site cannot be compared with a score from another.

Research gives a useful reference point. In a study of 270 adolescents using 3D imaging, whole-face symmetry averaged 53.49% (±10.73%) for males and 58.50% (±10.27%) for females. The authors noted that some differences were statistically significant without being clinically relevant.

Research benchmark: in 270 adolescents, average whole-face symmetry was 53.49% for males and 58.50% for females, with wide variation either side. Typical faces sit well away from perfect.

Those figures come from one method in one adolescent group, so they do not define a "good" score for you. They do show that ordinary, healthy faces vary widely and that perfect symmetry is not natural.

Symmetry is also only one ingredient in how a face is perceived. Psychological research links perceived attractiveness to a combination of averageness and symmetry, and averageness often predicts ratings more strongly for women, with symmetry contributing for men. Other work cautions that perfect symmetry is not natural and that online scores oversimplify facial harmony.

The sensible reading is this: treat a consumer score as a photographic estimate. If it is high, enjoy it. If it is lower than you hoped, it is not a verdict on your face.

Where can consumer symmetry tools go wrong?

The honest answer is in several places, and the literature gives us a clear picture of the size of the problem. A 2023 systematic review found that automated landmarking differs from manual landmarking by an average of 0.67–4.73 mm, and concluded that current automated methods are not yet accurate enough for clinical purposes. Note that the review examined 3D landmarking, so a 2D selfie tool should not be assumed to do better.

To put that range in context, the clinical linear asymmetries described in the adolescent study were often under 1–1.2 mm. A tool whose own error can match or exceed the size of the difference it is trying to detect is not measuring that difference reliably.

The most common sources of error:

Head yaw and tilt: turning or tipping the head even slightly moves the apparent midline and makes one side look larger.
Facial expression: a half-smile, a raised brow or a pressed lip changes the very landmarks being compared.
Lens distortion: a phone held close to the face enlarges the nose and centre of the face relative to the edges.
Low resolution or compression: blurred edges make landmark placement less precise.
Occlusion: hair, glasses and shadows hide or displace the points the algorithm needs.
Algorithm differences: definitions of landmarks vary between studies and populations, so two tools can disagree on the same photo.

Clinical 3D protocols deal with many of these problems. Landmark definitions based on curvature, standardised head positioning and surface-based alignment all reduce the dependence on a single viewing angle. The trade-off is that they need specialist equipment and trained operators.

Pro Tip: Upload the same photograph to two different tools, or take two separate photos on the same day. If the scores move by several points, the variation is telling you about the method, not about your face.

How do you get a cleaner result and keep your photos private?

You cannot remove the limits of the method, but you can remove avoidable errors from the photograph. A few minutes of preparation makes a visible difference.

Set up the camera at eye level. Use a tripod or rest the phone on a steady surface, and stand about an arm's length or more away. Close-up selfies exaggerate the centre of the face.
Face the lens squarely. Keep your chin level, your head straight and your shoulders relaxed. Check that both ears sit at a similar distance from the edge of the frame.
Keep a neutral expression. Relax your forehead and let your lips rest together without smiling.
Use even, diffuse light. Face a window or a soft lamp, and avoid light from one side, which creates shadows that look like asymmetry.
Clear the face. Choose a plain background, remove glasses, and tuck hair back so the brows, ears and jawline are visible.
Use a high-resolution file. A sharp JPG or PNG gives the software cleaner edges to work with.
Check how the tool handles your image. Find out whether processing happens in your browser or the photo is uploaded to a server, and whether it is stored. Where you have the choice, check the tool's privacy information before choosing in-browser processing.

Privacy deserves the same care as pose. A face photograph is personal data, so read the privacy notice before uploading and avoid tools that do not say what they do with your image.

When is facial asymmetry a medical emergency?

Longstanding asymmetry that you have always had is usually not a medical concern. A sudden change is different, and no online tool can tell the two apart.

The NHS uses the FAST test to spot a possible stroke or transient ischaemic attack (TIA):

Face: the face may have drooped on one side, the person may be unable to smile, or the mouth or eye may have dropped.
Arms: the person may be unable to lift both arms and keep them there because of weakness or numbness.
Speech: speech may be slurred or garbled, or the person may not be able to talk at all.
Time: call 999 immediately if you see any of these signs.

The NHS also lists further stroke warning signs, including blurred vision, severe headache and sudden dizziness. Do not wait to see whether symptoms pass, and do not photograph yourself to check first.

Not every sudden change is a stroke. Bell's palsy can cause one side of the face to weaken over hours or days, and facial palsy more broadly can affect eye closure and mouth movement. These need to prompt review by a GP or specialist, and the GP is the right first call when stroke signs are absent but the change is new.

Swelling is another reason to seek care. If one side of the face becomes swollen, this guide to facial swelling explains when dental or medical attention is appropriate.

Why do clinicians use 3D imaging and holistic assessment?

A clinician is not trying to produce a score. They are trying to understand your face well enough to advise you honestly, and a single symmetry percentage cannot do that.

In clinical research and planning, 3D cephalometric protocols align the skull or face to a stable reference using the basilar surface and iterative closest point methods, which helps estimate the symmetry plane with fewer landmark-related errors. Landmark-free surface matching goes further by comparing the whole shape of the face without relying on hand-placed points. This complexity exists because natural asymmetry is the norm, and a method must separate normal variation from a meaningful difference.

At our clinic, the assessment goes well beyond geometry:

We examine skin quality, volume, bone structure and the way the face moves.
We consider how each region relates to the others, not just to the midline.
We ask what bothers you and what you would like to change.
We recommend only what makes sense, and sometimes that is nothing.

A consultation is worth booking if a difference genuinely concerns you, if you are considering treatment, or if your online result has left you unsure what it means. If the change is sudden, go back to the previous section instead.

How does symmetry analysis fit with other facial harmony assessments?

Symmetry is one lens among several. On its own it describes balance between the left and right halves, but it says little about proportion, volume, skin condition or how the face looks in motion.

A fuller assessment combines it with other checks. Proportions look at how the upper, middle and lower thirds of the face relate, and how features such as the eyes, nose and lips sit in relation to one another. Contour assessment reviews the jawline, chin projection and cheek support. Skin health looks at texture, tone, hydration and signs of ageing, which strongly affect how a face is perceived.

Each of these can disagree with a symmetry score. A face can score well on symmetry yet look tired because of volume loss. Another can show visible asymmetry yet read as balanced and attractive because its proportions and skin are harmonious. This is why research on attractiveness treats averageness and symmetry as separate contributors rather than one measure.

In practice we consider the wider picture together with your own priorities. If you would like to see how a clinician reads that picture, our treatment overview shows the range of options we discuss, and our article on clinician-led AI in aesthetics explains where technology fits and where clinical judgement leads.

What are common asymmetry patterns and how are they usually corrected?

Almost every face has some asymmetry, but a few patterns come up often in consultation. Whether any of them calls for treatment depends on how much they bother you and on your anatomy.

Uneven lips. One side of the upper lip or one lip corner sits higher or fuller. A careful, conservative approach with lip filler can rebalance volume, though a smile-related difference may be better left alone.

Brow height difference. One brow sits higher or has a different arch, often because of habitual muscle use. Anti-wrinkle injections can relax an overactive muscle to even out position and expression.

Uneven cheek or mid-face volume. One side looks flatter. Volume can be restored with dermal filler, or skin quality can be improved with Profhilo or polynucleotides where quality rather than volume is the issue.

Jawline and chin difference. One side of the jaw looks less defined or the chin sits off-centre. Our guide to clinician-led jawline contouring explains how anatomy shapes the options.

Grooming and styling. Not every difference needs a procedure. A barber's guidance on shaping the beard cheek line shows how a simple change can alter how balanced the lower face looks.

Skeletal asymmetry, where the underlying bone differs, cannot be fully corrected with injectables, and we say so plainly during consultation.

How do we manage expectations about symmetry improvements?

Honest expectations are part of good care. The aim of aesthetic treatment is improved balance and natural-looking harmony, not mathematical symmetry, and it is worth being clear about that before anything is planned.

We explain several points to every client:

Perfect symmetry is neither achievable nor natural, and experts caution against chasing it.
Treatment can soften a visible difference, but it will not erase every one, and it may reveal another that was previously hidden.
Results depend on your anatomy, your skin and how your face moves, so we cannot promise a particular outcome.
Some treatments are staged over several sessions, and we review progress before adding anything.
An online score before and after treatment is a poor way to judge results, given the error built into those tools.

We would rather talk you out of an unnecessary treatment than give you one that does not serve you. A short, unhurried consultation lets us look at your face in person and in movement, which a photograph cannot do. Treatment prices are listed on our treatment pages, for example dermal fillers from £180, and a consultation will confirm what is right for you.

Our view: symmetry is one input, not a verdict

We see symmetry as more than geometry. The concern is whether something about a face is wrong. In our experience, a face that moves, expresses and looks like its owner reads as attractive far more reliably than one that scores well on a mirror test. A percentage cannot capture that, and it can easily make a perfectly healthy face feel like a problem.

We consider skin health, anatomy and your own goals. A consultation-led approach means we start by listening, look at your face as a whole, and recommend only what we would be comfortable recommending to a friend. If a difference is worrying you, or you have a clinical concern, we would be glad to see you for an assessment.

— Ms Sandra Petraskaite

This article is general information, not a substitute for advice from a qualified doctor. Consult a qualified healthcare professional about your own circumstances before acting on anything here.

FAQ

How can I test my facial symmetry?

You can upload a straight-on, evenly lit photo to an online symmetry tool, which estimates left-right balance from landmarks. For anything more reliable, a clinical assessment uses 3D imaging. Treat the online result as a rough guide, since automated landmarking differs from manual methods by 0.67–4.73 mm on average.

What's a good facial symmetry score?

There is no single good score, because tools define and weight symmetry differently. In one 3D study of adolescents, whole-face symmetry averaged 53.49% for males and 58.50% for females. Compare scores only within the same tool, and do not read them as a judgement on your appearance.

Can ChatGPT analyze my face shape?

A general AI chatbot can describe a face shape from a photo, but it is not built to measure symmetry or landmarks precisely, and its description is subjective. It cannot give a clinical measurement. If privacy matters to you, check what happens to any photo you upload before sharing it.

Is anyone's face 100% symmetrical?

No. Natural asymmetry is the norm, and researchers caution that perfect symmetry is not natural. Clinical studies of healthy people show clear left-right differences. A sudden new asymmetry is different, and the NHS advises calling 999 for signs of a stroke.

Sources

Ready to explore your options?

Speak with a NexGen clinician for a personalised consultation. We’ll listen, assess, and recommend the safest and most effective route — always led by evidence, always tailored to you.

Book a consultation or get in touch with our Cambridge team.