What Skin Analysis Actually Shows — and What It Cannot Tell You

Skin analysis means photographing and examining the skin under several different kinds of light — ordinary white light, polarised light and ultraviolet — to reveal what the naked eye misses: pigment sitting below the surface, dilated vessels, dehydrated zones, and sebum inside pores. Each type of light penetrates to a different depth, so the same face produces completely different "maps" depending on the mode. But one thing needs saying up front: skin analysis is not a diagnosis. It produces images and numbers; interpreting them and reaching a conclusion remains a doctor's job.

This guide explains what each lighting mode genuinely shows, which figures in your report are measured and which are merely estimated from a photograph, how far AI helps and where it breaks down — particularly on Vietnamese skin — and the questions worth asking once you see your results.

Written by the dermatology team at Eternal Beauty Center (An Hội Đông Ward, formerly Go Vap District, Ho Chi Minh City), drawing on DermNet NZ, the Journal of Clinical and Aesthetic Dermatology, research indexed on PubMed/PMC, and Nature Medicine.

1. What skin analysis is — and what it isn't

Skin analysis is an imaging technique, not a test. Getting that distinction right determines how you should use the results.

A modern system has three parts: a light source that switches between wavelengths and polarisation states; an imaging enclosure that fixes distance, angle and ambient light so sessions are comparable; and software that measures, marks abnormal areas and outputs a table of scores.

Three things it does well:

• Sees below the surface — deep pigment, vessels, lesions not yet visible on the skin.
• Quantifies and compares over time — counting lesions, measuring pigmented area, re-photographing after eight weeks under the same conditions to see whether a protocol is working.
• Lets the patient see what words struggle to convey — in practice, its single biggest value in consultation.

Three things it cannot do:

• Diagnose disease. Images suggest directions, they don't conclude. Melasma, seborrhoeic dermatitis, rosacea and discoid lupus can all produce overlapping appearances.
• Replace examination and history. What medication you take, your hormonal status, what you've already tried — no camera captures that.
• Screen for skin cancer. That requires a dermatologist examining you directly, purpose-built dermoscopy, and a biopsy where indicated.

The line worth remembering: skin analysis answers "what does this skin look like at different depths". It does not answer "what condition does this skin have".

2. Why changing the light changes what you see

Light hitting skin reflects in two places: part of it bounces straight back off the surface, and the rest travels deeper before scattering back out. The whole technique rests on separating those two — keep the surface reflection to study texture, remove it to look underneath.

Mode	Principle	Shows best
Plain white light	An image close to what the eye sees; the baseline for comparison	Overall tone, obvious lesions, inflammatory acne
Parallel-polarised	Retains surface reflection	Surface: scale, texture, wrinkles, pores, oil shine
Cross-polarised	Blocks surface reflection, keeps only light returning from depth	Subsurface: pigment, vessels, erythema — glare removed, so contrast rises sharply
Ultraviolet (Wood's lamp)	UV around 365nm, observing the fluorescence emitted	Pigment borders, dehydrated zones, bacterial fluorescence
Raking light	Angled from one side to cast shadows	Surface topography: atrophic scars, roughness

The two polarisation modes are the pair worth understanding. Parallel-polarised is like looking at a lake in harsh sun — you see every ripple on the surface and nothing beneath it. Cross-polarised is like putting on polarised sunglasses — the glare disappears and you can see the bottom. The literature records cross-polarisation as probably the best method for visualising hyperpigmentation, while also enhancing erythema and increasing lesion-to-skin contrast.

Which means: a skin-analysis photograph tells you nothing unless you know which mode produced it. On the same cheek, the parallel-polarised image can look fine while the cross-polarised one reveals a large field of pigment.

3. What it shows about pigment — and where people over-read it

For anyone with melasma, the important question isn't "is there pigment" but "at what depth does it sit" — because that governs treatment time and realistic expectations. It is also where skin analysis gets over-interpreted most often.

The classic rule under a Wood's lamp:

• Epidermal pigment (upper layer): borders appear sharper and darker than to the naked eye — UV raises contrast in the superficial layer.
• Dermal pigment (deeper): not accentuated, appearing grey-blue with indistinct borders.
• Mixed type: some areas accentuate and others don't — and in practice this is the most common pattern.

Here's the part rarely mentioned: this classification is contested in the literature. Some studies report good correlation with histology; others indicate that essentially all melasma has a dermal component — which makes "yours is superficial, it'll clear quickly", based on a UV photograph alone, a conclusion that outruns the evidence.

How to use the information properly:

• Treat the Wood's-lamp finding as one input, weighed alongside age of onset, hormonal factors, treatment history and actual response — not as a verdict.
• Be wary of consultations promising a precise treatment timeline from a single UV image.
• What genuinely changes outcomes is not the superficial/deep label but controlling the triggers — sun, heat, hormones. See melasma: causes and treatment.

Skin analysis also helps separate pigmentary conditions that are easily confused — melasma, freckles and sun-induced brown spots differ in distribution and border; see telling melasma, freckles and brown spots apart. Where intervention is warranted, the direction is weighed case by case between Mela Peel for post-inflammatory marks and E-Mela for resistant melasma.

4. What it shows about acne and sebum

**Under UV, the orange-red fluorescent dots appearing inside pores are porphyrins — produced by Cutibacterium acnes within sebaceous follicles.** It is the most striking image in a skin-analysis session, and the most frequently misread.

What the orange dots do mean: there is activity from resident bacteria in sebaceous follicles, usually alongside higher sebum output — most often across the T-zone.

What they do NOT mean:

• They don't measure acne severity. C. acnes is a normal resident organism on healthy skin. Plenty of orange dots doesn't equal severe acne, and severe inflammatory acne can present with little fluorescence.
• They are not something to "eradicate". Acne isn't a simple infection — it's the combination of follicular hyperkeratinisation, sebum, bacteria and inflammation. See what acne is.
• They aren't specific. Coral-red fluorescence in skin folds usually comes from a different organism (Corynebacterium) causing erythrasma, unrelated to acne.

The real value of UV mode for acne-prone skin lies elsewhere: a sebum map showing where oil concentrates so products can be targeted, and dehydrated zones hidden under oily skin — skin that is both shiny and water-deficient is routinely mismanaged because only the oil is visible.

Findings typically inform the choice between medical-grade extraction (O2 Peel) and a broader acne treatment protocol where deeper inflammation is present.

5. What it shows about the skin barrier

Skin analysis cannot measure barrier function directly, but it does pick up indirect signs — and this is the group of scores most prone to being oversold.

Observable signs:

• Background erythema on cross-polarised images — rises when the barrier is compromised and inflammation persists.
• Fine scale and coarse texture on parallel-polarised images.
• Dehydrated zones appearing as darker patches under UV.
• More visible superficial vessels, sometimes suggesting thin skin or a vascular condition.

One distinction matters: the "hydration" figure in your report measures surface conductance of the stratum corneum — it is not TEWL (water lost through the skin), which is the direct measure of barrier function and requires dedicated instrumentation under standardised conditions. What both mean, and how to repair a compromised barrier, is covered in the skin barrier guide.

Practical consequence: if a session shows diffuse erythema and coarse texture, the right move is usually to postpone invasive treatment for two to four weeks and repair first — not to start something strong immediately.

6. Which numbers are measured and which are estimated

Not every figure in the report carries the same weight — and nobody tells you which is which. This is the most important section of the article.

Metric group	Nature	Reliability	Reading it
Lesion counts (acne, spots)	Counted on the image	High	Comparable across sessions if conditions match
Pigmented area / borders	Measured on the image	High	The best progress marker for melasma
Surface texture, pores	Measured on raking/parallel-polarised images	Fair	Sensitive to lighting and angle
Erythema / vessels	Colour-channel analysis	Fair	Shifts with room temperature, recent exercise, recent cleansing
Surface hydration	Conductance/capacitance	Moderate	Strongly dependent on room humidity, temperature and time since cleansing
Elasticity, "firmness"	Estimated from an image	Low	Not a true mechanical measurement
"Skin age"	Algorithmic composite score	Low	A communication figure with no agreed clinical standard
Future ageing prediction	Illustrative simulation	Not for decisions	A visualisation, not a medical prognosis

Three self-defence questions when a consultation leans on scores:

1. Is this measured, or estimated from a photo?
2. Was the previous session shot in the same room conditions and at the same time of day?
3. If this number improves, what visible change should I see on my skin?

If a metric can't answer question three, it shouldn't be the basis for spending money on a treatment.

7. AI in skin analysis: what works and what fails

AI is genuinely good at counting, outlining and comparing between sessions. It is not reliable at concluding what a condition is — and it is less accurate on darker skin.

Where AI actually helps:

• Automatically outlining and counting hundreds of lesions faster and more consistently than a person.
• Registering images across visits to surface the change — something human eyes are poor at.
• Standardising scoring between different operators.

Two limits to know before trusting a result:

First — skin-tone bias. Most models are trained predominantly on lighter skin (Fitzpatrick I–III). A 2022 Nature Medicine study recorded sensitivity dropping by up to 34% on Fitzpatrick V–VI skin, and independent evaluations show accuracy declining as skin tone darkens. Vietnamese skin typically falls in Fitzpatrick III–IV — inside the range where models start to lose reliability.

Second — the gap between marketing figures and real-world results. Vendors commonly cite 69–95% agreement with dermatologists on cosmetic grading. But in independent testing on real patients with common skin conditions, top-diagnosis accuracy for several apps landed around 47–59%. Furthermore, no consumer AI skin app is cleared for diagnosing skin cancer.

The practical conclusion: use AI as a measuring and tracking assistant, not as a diagnostician. Every conclusion should pass through a doctor.

8. What a session involves — and how to prepare

A proper session takes 15–30 minutes and always ends with a doctor interpreting the results — if all you get is photographs and a scorecard, that's a sales process, not an examination.

The usual sequence:

1. Cleansing and makeup removal — makeup, sunscreen and some actives all distort UV fluorescence.
2. A few minutes to stabilise — coming straight in from the sun skews erythema and hydration readings.
3. Capture across each lighting mode, with a chin rest fixing the angle.
4. Software analysis, producing the score table and comparison images.
5. The doctor examines you directly and takes a history, cross-checks against the images, and only then concludes.

Preparing so the results mean something:

• Arrive without makeup; if you're wearing it, come 10 minutes early to remove it properly.
• No exfoliation, peels or microneedling for 7 days beforehand — actively flaking skin distorts texture and erythema.
• Avoid significant sun exposure for 48 hours.
• For ongoing tracking, keep the same time of day and the same point in your menstrual cycle between sessions.
• Bring a list of the products and medications you use — this shapes the conclusion more than any score does.

On how often to re-assess and why periodic review beats a single scan, see why regular skin monitoring matters.

9. 5 questions to ask when you get your results

Results only matter if you understand what decision they lead to. These five turn a scorecard into a plan:

1. "Of all these metrics, which is actually my main problem?" — Forces prioritisation. Ten red scores don't mean ten treatments.
2. "Is this measured, or estimated from the photo?" — See section 6.
3. "What happens if I do nothing for three months?" — Separates what needs attention from what is purely cosmetic.
4. "What's the cheapest, least invasive thing to try first?" — Often correct sun protection and barrier repair are enough to start.
5. "When do we re-assess, and what would tell us it's working?" — Without a checkpoint, you can't tell whether a protocol is doing anything.

10. When skin analysis isn't enough

Some situations aren't just beyond what imaging can settle — stopping there can cause a dangerous delay.

See a dermatologist in person, and don't rely on a machine, for:

• A mole that changes in size, shape, colour or border; a lesion that bleeds or won't heal.
• A new lesion appearing after age 40 and growing.
• Clustered blisters, weeping, yellow crusting, escalating pain — suggesting infection.
• Persistent facial redness with visible vessels and flushing episodes — possibly rosacea, which needs its own pathway.
• Patchy hair loss or lesions with scarring atrophy — autoimmune conditions requiring proper assessment.

Any lesion suspicious for malignancy needs purpose-built dermoscopy and a biopsy where indicated — an entirely different discipline from cosmetic skin imaging.

Medical disclaimer: this article is educational and does not replace a consultation. Accurate diagnosis requires a dermatologist examining the skin directly and taking a history.