Has anyone tried Visia skin analysis or Pharmanex® BioPhotonic Scanner and reversed some of their metrics?

#1

I tried it. It showed some wrinkles below my eyes where the skin is thinner (even though my skin appears perfect from the outside), which might be related to not enough collagen…

Essentially 0 UV damage/age spots/most other types of damage.

I also have porphyrins, but it’s unclear if this is aging-related so this is not too concerning…

===

Your Skin Carotenoid Score Revealed Now

Dear Alex,

Your Pharmanex® BioPhotonic Scanner results are in! Your personalized scan, measured by X on 12/15/2023, returned a Skin Carotenoid Score (SCS) of 69000.

Your Skin Carotenoid Score is an immediate measurement of your own skin carotenoid content and an important indicator of your body’s antioxidant defense system. The higher your score – the more carotenoids are present in your skin.

This was 99th percentile

3 Likes
#2

I have uaed visia scan and improved some metrucs, but i am currently missing some data from hube 2023 that i need to do a proper analysis.

#3

This is bryan johnson’s “results”

MARKERS

Spots: Age 10
Wrinkles: Age 10
Texture: Age 14
Pores: Age 50
UV Spots: Age 62
Brown spots: Age 11
Red areas: Age 70 
Autofluorescence: age 32
Multispectral imaging face age: 41

Obviously looking at his skin, he doesn’t look that young [this is really hard for him to rejuvenate], and I suspect it’s structural (elastin/collagen) stuf more so than UV spots or age spots… This at least gives hopes that wrinkles are reversible

2 Likes
#4

I don’t know how to link the count of each aspect (and I don’t have autofluorescence or Multispectral) to an age.

#5

Alex, where did you get this done?
I’ve called some places in the SF Bay Area that mention Visia analysis on their website, but then I call and they’ve phased it out and no longer offer it.

#6

I would think it is relatively old tech. It is, however, a system where you can objectively compare facial skin over a period of time. I have done the measurements about 6 times and I intend analysing the change in a spreadsheet as soon as I get all my June 2023 results.

I went on holiday over the summer and because I was concerned that I may not be able to maintain my protocol at its highest level whilst on holiday I reduced it some weeks before going on holiday. Whilst on holiday I was in the sun as well.

What appears to have happened is that the skin quality deterioated a bit on holiday and has since been recovering, but I need to do the analysis.

Here is an example comparison.

20231206_122035
20231206_1220351920×1440 303 KB

4 Likes
#7

I haven’t tried the Visia scanner but I tried the Pharmanex Biophotonic scanner in 2004 and got scores of 41000 and 38000. AFAIK it’s mostly serves as an indirect indictor of carotenoid intake and deposition in skin.

1 Like
#8

I have just done another Visia scan and attach one of the comparative photos. Sadly I have to photograph the computer that runs it with my mobile which is not ideal.

The left is February 2024 and the Right May 2025.

20250519_111216
20250519_1112161920×1440 243 KB

I find skin something useful to compare because it is all about the maintenance of the Extra Cellular Matrix. My assumption is that given that I am not doing anything specific to the skin beyond exposing it to the sun without sunscreen and using soap and water (ie no botox), if I can get the structure of the
ECM to improve then I can assume similar activity is likely in other tissues.

The Visia Scan’s analysis is a little unpredictable I think, but at least I have photographs in similar lighting to compare which is useful. In a broader sense I think the direction of travel involves reasonably good maintenance of the ECM.

2 Likes
#9

I just realized that bc it’s undertrained on teens (like grimage), visia really overestimates the skinage of teenagerS

o3: (way better than g5t)

Short version: the camera isn’t broken, the database is.

VISIA’s “world-largest” skin dataset starts at 18 years old. The marketing brochure brags about an “Advanced Aging Simulation 18–80” and never mentions anyone younger. Teen skin simply isn’t in the reference pool.

Academic validation studies echo that bias. The big accuracy paper Canfield likes to wave around tested 117 adults, age 18-89; mean error was only ≈1.4 years… in adults. No published VISIA study I could find includes minors.

Result: teenagers get shoved up the age scale. When the algorithm sees hormonal acne, enlarged sebaceous pores or early sunspots it tries to map them to the nearest pattern it has—late-20s or 30-something faces. A 15-year-old with breakouts can pop out as “TruSkin 28,” while a porcelain-smooth 17-year-old might read “TruSkin 19.” Anecdotal reports of 16-year-olds being told they look 32 are all over user forums.

So yes, VISIA often overshoots teen age because:

  1. Training gap: No under-18 cohort → poor calibration.
  2. Metric design: Wrinkle/UV-spot scoring is weighted toward adult photo-aging markers, not hormonally inflamed teen skin.
  3. Percentile comparison: Scores are normalized to “peers” that don’t exist for a teenager, so the percentile math collapses.

If a clinic tries to sell a 14-year-old an anti-aging laser based on a VISIA printout, remind them the machine literally has never met a 14-year-old.

G5T …

Short answer: usually, yes.

Why it happens:

The reference database skews adult. If teens are rare in the training/“norms,” the model extrapolates from grown-up skin and gets jumpy.

Teenage skin is a paradox: low wrinkles but high sebum, porphyrins, redness, and larger-looking pores. Adult-trained metrics read those as “damage” and bump the age.

Acne, post-inflammatory marks, and uneven tanning light up UV/spot features that the system weighs as aging in adults.

If your phototype or ethnicity is underrepresented, expect even spicier errors.

How to use it without losing your mind:

Ignore the single “skin age” number. Track feature percentiles over time (wrinkles/texture > spots/porphyrins for teens).

Re-scan under identical lighting and prep, no active breakouts or heavy SPF cast.

If the device allows age-matched comparisons, use those; if not, treat results as noisy vibes, not diagnostics.

So yes: teen skin gets penalized for being alive and full of hormones while the algorithm desperately tries to fit you into a 35-year-old’s rulebook.