Topic

Taking a good hair photo

A well-taken photo shows curl family, length, hairline shape and overall hair loss pattern, but it cannot show strand thickness, porosity or scalp oiliness, and small changes are invisible.

hair loss receding hairline thinning low volume frizz dullness split ends breakage gray hair dandruff poor curl definition

Key takeaways

Hair state matters more than the camera. Clinical protocols require hair that has been washed, dried and combed, with no product. Wet hair clumps and reflects light, so it shows more scalp and makes hair look thinner than it is Rubin 2023. Product, brushing, heat styling and humidity all change how curl looks Wortmann 2010, Gao 2007.
Hair-loss staging from photos is only moderately reliable, even for experts. Dermatologists grading Norwood stage from photos reached an intraclass correlation of only 0.63-0.68, and their repeat ratings were poor Guarrera 2009. The BASP scale was more reproducible than Norwood-Hamilton Hong 2013. Think in broad bands rather than a precise stage.
Men are only moderately good at rating their own balding. Against trained observers, self-ratings agreed exactly 48-55% of the time (kappa 0.39-0.46), while two trained observers agreed 85% of the time Taylor 2004, Littman and White 2005. A standardized photo set is a better anchor than memory.
Small changes are invisible. In standardized scalp photos, a 205-person group needed a density difference of about 22.8% (41.7 hairs/cm2) to tell two scalps apart, no different from dermatologists Oh 2025. Month-to-month photos will not show early loss or early regrowth.
Strand thickness, porosity, elasticity and true density are invisible in normal photos. Measuring diameter or hairs per cm2 needs magnified scalp imaging such as trichoscopy or a phototrichogram Pyo 2026, Majd 2021.
Shine in a photo is mostly about the lighting. Measured luster depends on light geometry, viewing angle, hair color (darker hair shows more contrast), fiber shape and curl, and surface oil Keis 2004, McMullen 2004. Even lab imaging struggles to quantify small shine changes Kaplan 2009.
Curl subtype is hard to read even in the lab. Trained raters measuring single fibers on templates agreed only modestly on 8 groups (kappa 0.42), and among 48 volunteers none had all 10 sampled hairs fall into the same curl group Mkentane 2017.
Close-range selfies distort the face. A smartphone held about 20 cm (8 inches) from the face stretched midface proportions by about 18% Derakhshan 2024, Ward 2018. Applying this to hairline photos is our inference, but take hairline photos from at least arm's length, ideally by a helper, with the main (1x) lens.
AI accuracy figures are mostly lab results. A 2026 systematic review of machine learning in androgenetic alopecia found 17 studies; only 3 were validated in clinical settings, 47% had high risk of bias, and some skin tones and ethnic groups were under-represented Huang 2026.

Why hair photos mislead

Hair state. Wetness, product, days since washing, brushing, heat, water setting and humidity all change what the camera sees. Clinical photography uses hair that has been washed, dried and combed without product, and excludes wet hair because it increases reflection and exposes more scalp Rubin 2023. Trials also keep the same hairstyle, color, parting and combing at every visit Dhurat 2009.
Light. Direction, softness and color of light change how much scalp shows, how shiny hair looks and what color it appears. Trial photography fixes camera and flash on a positioning device so view, magnification and lighting are identical each visit Dhurat 2009, Canfield 1996. Consensus imaging standards cover lighting, background, field of view, focus, resolution and color calibration Finnane 2017.
Camera and lens. Short camera-to-face distances enlarge features nearest the lens Derakhshan 2024.
The viewer. People judge 3D hair poorly from flat images Daniels 2021. A hair-volume difference of about 3 cm2 on a swatch image was the just-noticeable difference, and 2 cm2 could not be discriminated Muthunilavan 2026 (industry lab).

What a photo can show, trait by trait

Trait	Can a photo show it?	Main distortions and best conditions
Curl family (straight, wavy, curly, coily)	Partly: reliable for the broad family only on clean, product-free, air-dried hair	Wetness, gel, brushing, heat or water setting, humidity and length change the pattern; different zones differ. Use back and side views plus close-ups of crown, nape and temples De la Mettrie 2007, Wortmann 2010, Callender 2026
Curl subtype (A/B/C)	Not reliably	Even trained raters agree only modestly; curl varies between hairs on one head Mkentane 2017, Loussouarn 2007
Length	Reliable for straight and wavy; partly for curly and coily	Shrinkage hides length; ask for stretched length too
Strand thickness	No	Individual hairs are tens of micrometres wide, far below what a head-and-shoulders photo resolves Pyo 2026, Majd 2021
Apparent density and scalp show-through	Partly	Wet hair clumps; light direction, flash, hair-skin contrast and parting matter. Same parting and lighting each time Rubin 2023, Rose 2025
True density and miniaturization	No	Needs trichoscopy or a phototrichogram, and even automated counting makes errors Saraogi 2010
Hairline and pattern-loss stage	Partly: broad stages yes, neighbouring stages unreliable	Fringe combed forward, head tilt and lens distortion matter Guarrera 2009, Hong 2013
Change in density over time	Partly	Differences below about 20% are not reliably noticed, even in standardized photos Oh 2025
Volume	Partly	Days since wash, product, sebum and curl all matter Harrar 2026
Frizz	Partly	Humidity, time since styling and touching matter; photograph against a plain background Gao 2007, Skedung 2021
Shine	No (not reliably)	Light geometry, angle, hair color, curl and sebum all change it Keis 2004, McMullen 2003, Lefaudeux 2009
Visible damage	Partly	Uneven, see-through ends and short broken hairs show; shaft defects need magnification Rudnicka 2018, Rudnicka 2013
Gray proportion	Partly: good in dark hair, weak in blond or light hair	A published graying score samples five scalp sites Singal 2016
Flaking	Partly	Trials use trained in-person graders scoring 8 scalp sections Bacon 2014 (industry)
Porosity, elasticity, oiliness	No	These are physical and chemical properties; describe your history instead (see hair porosity)

Hair loss photography

Global photography was introduced to hair-growth research in 1987 and became central to the finasteride trials, where an expert panel reviewed photographs alongside hair counts Kaufman 1998, Rubin 2023. The method uses four standard views (vertex, midline, frontal and temporal), a fixed head position and a 7-point scale from "greatly decreased" to "greatly increased" Dhurat 2009, Olsen 2012. A 2025 UK consensus guide proposes protocols for pattern loss, frontal fibrosing or traction alopecia, and patchy loss Takwale 2025. A 2025 US clinic project defined a minimum of three views (frontal, both temples, crown) and noted that light hair on fair skin gives poor contrast and curly or braided hair may need different positioning Rose 2025.

How reliable is grading from photos? Moderate at best.

Seven dermatologists and 16 residents grading 43 photos reached ICC 0.63-0.68 on Norwood-Hamilton; re-grading three months later agreed poorly with their own first ratings. The authors concluded the scale is reliable only if collapsed into broader classes Guarrera 2009.
Forty raters grading 100 photo sets matched each other more often with BASP basic types than with Norwood-Hamilton Hong 2013.
For women, the Sinclair midline-part scale correlates well with a trichoscopy density measure, making the part-line photo the most informative single view Kasprzak 2019.
Self-ratings are repeatable but biased: consistent on retesting (kappa 0.71-0.81) but only moderate against an interviewer (kappa 0.47) Littman and White 2005. That favors photos plus questions.

How much change is visible? In 205 people (100 with alopecia) viewing standardized photos with known counts, the just-noticeable difference was about 41.7 hairs/cm2, a 22.8% change, not significantly different from dermatologists or between people with and without alopecia Oh 2025, Buontempo 2024. That figure is for spotting a difference between two photos side by side, not for noticing thinning in everyday life. In 56 PRP patients, self-reported improvement tracked measured density but not hair width Brinks 2025. Hair mass depends on both density and diameter, and diameter is invisible in ordinary photos Wikramanayake 2012.

What only magnification shows: hair counts, diameters, miniaturized hairs and the anagen-to-telogen ratio need trichoscopy or a phototrichogram Dhurat 2009, Chamberlain and Dawber 2003. One automated system overestimated total density and missed hairs that crossed or shared a follicle opening Saraogi 2010. A smartphone with a dermatoscope lens matched manual trichometry within a 20% margin, taking 24 seconds of clinician time versus 9.2 minutes Majd 2021.

Curl in photos

Lab curl typing measures single fibers, not heads. The segmentation-tree method classifies hair into eight types using curve diameter, curl index (stretched divided by relaxed length) and number of waves. It was built from 1,442 people in 18 countries, then 2,449 people in 22 countries De la Mettrie 2007, Loussouarn 2007 (industry). Fibers are measured on 6 cm of hair near the root, after washing and air-drying for at least 4 hours Mkentane 2017. Software can also measure curvature from photos of strands Lasisi 2016.
Rater agreement is modest even in the lab: kappa 0.42 for 8 groups and 0.67 for 6 groups, and worst at the extremes Mkentane 2017.
Curl varies within one head, so photos should sample several zones Callender 2026.
Water, heat and humidity reset shape. Modelling suggests about 50% recovery of a wet-set shape at 65% relative humidity after roughly ten times the original setting time Wortmann 2010. Heat-shaped and chemically shaped fibers relax by a common mechanism Breakspear 2024. A photo therefore shows recent history (how hair was dried and styled, and the air humidity) as much as the natural pattern.

Volume, frizz and shine

Volume: a 2026 industry method photographing the head from several angles produced root-lift measures that correlated with expert and consumer ratings (about 0.65-0.88) in 40 women Harrar 2026. Under high humidity, European-origin tresses swelled more than East Asian tresses Cruz 2017 (tresses only).
Frizz: lab frizz is measured by photographing tresses at controlled humidity and tracking changes in outline and volume; naturally curly and kinky hair was the frizziest Gao 2007. Rising humidity raises perceived frizz in curly hair partly by changing friction Skedung 2021 (industry).
Shine: luster depends on the contrast between mirror-like and diffuse reflection. Darker hair absorbs more scattered light, so it looks shinier Keis 2004, McMullen 2003. Fiber diameter, cross-section and curvature broaden the highlight and lower luster Keis 2004b. Tight curls break the highlight into scattered points, and synthetic sebum reduces luster McMullen 2004. Polarization imaging tracked the ratings of 15 untrained panelists Lim 2006 (industry lab). Because a phone photo fixes none of these factors, it cannot fairly compare shine between people, only one person under identical conditions.

Damage, gray hair and the scalp

Damage: smooth texture, shine and clean-cut ends are read as healthy Sinclair 2007, and uneven or see-through ends show in a close-up. Under the microscope, abnormal shaft findings were seen in 17% of untreated controls versus 46-54% of women who dyed, permed or flat-ironed Talghini 2013. In bleached or poor-quality hair, splits start inside the strand and run further Taylor 2024. Damage history is better captured by questions.
Gray hair: a graying severity score sampling five scalp sites had very high rater agreement (ICC 0.967), suggesting zone-based gray estimates from photos are feasible in dark hair Singal 2016. Single-hair scanning showed that some gray hairs naturally regain pigment Rosenberg 2021.
Flakes and redness: photos can flag visible flaking, but diagnosis belongs to a clinician.

AI on hair photos

17 machine-learning studies on androgenetic alopecia: only 3 clinically validated, 47% at high risk of bias Huang 2026. Fifteen AI trichology studies were mostly retrospective and single-source Alanazi 2026.
Male pattern grading from clinical images: 761 images from 257 patients gave 96-98% average precision for outlining loss regions, but accuracy dropped for intermediate grades Xi 2025.
Trichoscopy images: separating early androgenetic alopecia from none reached 0.90 accuracy, but on a hold-out set of only 20 images Suh 2025. Automated phototrichogram density correlated r = 0.73 with clinicians, with wide limits of agreement (-53 to +29 hairs/cm2) Pyo 2026.
Smartphone photos of shed hairs: counts were within about 5 hairs (7%) of manual counts and beat visual estimates, validated only for dark hairs on light backgrounds Zhang 2026.

What it means for your hair

Straight and wavy hair: curl family reads well. Waves are easily erased by brushing, blow-drying or weight and boosted by product, so wavy hair is the type most often mislabelled from styled photos (our inference from Wortmann 2010). Scalp show-through is visible earlier in fine, straight, dark hair on pale skin.
Curly and coily hair: shrinkage hides length, curl differs across zones and subtypes are hard to call even in the lab Mkentane 2017. Protective styles hide density and curl pattern. Clinics are advised to position textured hair differently and ask patients to come with hair worn as naturally as possible Rose 2025. Coils cover the scalp, so photos may understate early thinning, and part-line photos need careful sectioning.
Strand thickness cannot be read from photos. Fine hair makes thinning look worse and coarse hair masks early loss, because appearance depends on hair mass, not count alone Wikramanayake 2012.
Density gives only a rough low, medium or high impression; see hair density.
Hair-skin contrast: light hair on fair skin under-shows thinning Rose 2025; dark hair on pale skin over-shows it.

Myths and facts

"Wet-hair photos show your real density." False: wet hair clumps and reflects light, revealing more scalp Rubin 2023.
"Shiny hair in a photo means healthy hair." Not reliably: shine depends on light, angle, color and processing Keis 2004.
"My curl type is what I see in my styled photos." Product, drying method and humidity change the pattern, and curl varies across the head Wortmann 2010, Mkentane 2017.
"An app can give my exact Norwood stage from a selfie." Not demonstrated: experts disagree between neighbouring stages Guarrera 2009 and AI studies have limited external validation Huang 2026.
"Photos are pointless for hair loss." False: standardized photos remain a core trial outcome Olsen 2012 and beat comparing against memory Dhurat 2009.

How to take a good set of hair photos

This protocol adapts clinical practice Dhurat 2009, Rubin 2023, Rose 2025, Finnane 2017 for home use. The specific distances and lighting choices are practice-based, not trial-tested.

Prepare your hair on wash day. Wash with your usual shampoo (conditioner is fine). Use no leave-in, gel, cream, oil, dry shampoo or spray. Let hair dry completely, ideally air-drying for several hours (lab curl protocols use at least 4 hours). Do not brush or scrunch once dry. If you have braids, locs, extensions or a wig, take photos with hair out if possible.
Set up the light. Face a window in daytime for soft, even light, or use bright, even room light. Avoid a single overhead spotlight, direct sun and mixed light colors. Turn flash off. Stand in front of a plain wall that contrasts with your hair color.
Set up the camera. Use the rear camera if you can, with the main 1x lens (not ultra-wide), and turn off filters, beauty mode and portrait blur. Keep the camera at least 50 cm (about arm's length) from your head; a helper at about 1 m is better. Hold it level with the area and point it straight at it. Use the timer and a stand if no one can help.
Take the hairline and hair loss set: the front with hair pushed straight back off the forehead (neutral face, camera at eye level); left and right temples with the head turned about 45 degrees; the top from directly above with a straight centre part from the front hairline to the crown (the most important view for women); and the crown from above and slightly behind.
Take the hair type set: back and side views with hair down; a close-up of mid-lengths and ends held against plain paper (white for dark hair, black for light hair); and optionally 3-5 naturally shed hairs laid on paper next to a ruler, which shows curl shape and length but not thickness.
Check each photo is in focus with the whole area in frame, no glare across the scalp, no shadow over the part, no motion blur and no filter.
To track over time, repeat every 3 months in the same place, light, camera, parting and days since washing, and compare side by side. Changes smaller than about 20% will not be visible yet Oh 2025. Noticeable results from treatment often take more than 6 months Rose 2025, and the pivotal finasteride trials judged photos at 1 and 2 years Kaufman 1998, Olsen 2012.

Remember that photos give broad bands, not exact grades. Confirm anything marked partly or no in the table with questions (see how to work out your hair type at home), and see a dermatologist for patchy loss, redness, scale or pain rather than relying on a photo.

Bottom line

A careful photo on clean, dry, product-free hair is a good record of curl family, length, hairline and coarse changes, and a poor tool for thickness, porosity, shine, oiliness or small changes. Take the same set in the same conditions every few months and compare side by side. Apps and AI reads are best treated as a rough first guess.

Studies 56

Rubin A, Shah RR, Khan S, Haroon A, Rao B. Imaging tools for noninvasive hair assessment. Cutis. 2023;112:E52–E57

Design: review

Result: Clinical hair photography requires washed, dried, combed hair with no product, since wet hair clumps and reflects light, exposing more scalp.

Key takeaways

Why hair photos mislead

What a photo can show, trait by trait

Hair loss photography

Curl in photos

Volume, frizz and shine

Damage, gray hair and the scalp

AI on hair photos

What it means for your hair

Myths and facts

How to take a good set of hair photos

Bottom line

Studies 56

Rubin A, Shah RR, Khan S, Haroon A, Rao B. Imaging tools for noninvasive hair assessment. Cutis. 2023;112:E52–E57

Dhurat R, Saraogi P. Hair evaluation methods: merits and demerits. Int J Trichology. 2009

Canfield D. Photographic documentation of hair growth in androgenetic alopecia. Dermatol Clin. 1996

Takwale A et al. A practical guide to the standardization of hair loss photography for clinicians. Clin Exp Dermatol. 2025

Rose L et al. Standardization of clinical photos for tracking management of hair loss in dermatology clinics. J Cosmet Dermatol. 2025

Finnane A et al. Proposed technical guidelines for the acquisition of clinical images of skin-related conditions. JAMA Dermatol. 2017

Kaufman KD et al. Finasteride in the treatment of men with androgenetic alopecia. J Am Acad Dermatol. 1998

Olsen EA et al. Global photographic assessment of men aged 18 to 60 years with male pattern hair loss receiving finasteride 1 mg or placebo. J Am Acad Dermatol. 2012

Guarrera M et al. Reliability of Hamilton-Norwood classification. Int J Trichology. 2009

Hong H et al. Reliability of the pattern hair loss classifications: a comparison of the basic and specific and Norwood-Hamilton classifications. J Dermatol. 2013

Taylor R et al. Validity of self reported male balding patterns in epidemiological studies. BMC Public Health. 2004

Littman AJ, White E. Reliability and validity of self-reported male balding patterns for use in epidemiologic studies. Ann Epidemiol. 2005

Kasprzak M, Sicińska J, Sinclair R. The Trichoscopy Derived Sinclair Scale. Australas J Dermatol. 2019

Oh CS et al. Evaluating patient's ability to detect changes in hair density using standardized global scalp photography: a cross-sectional study. J Am Acad Dermatol. 2025

Buontempo MG et al. Evaluating subjective versus objective measures of changes in hair density using standardized scalp photography. J Am Acad Dermatol. 2024

Wikramanayake TC, et al. Cross-section Trichometry: A Clinical Tool for Assessing the Progression and Treatment Response of Alopecia. Int J Trichology. 2012

Brinks A et al. Evaluating the accuracy of patient-reported hair outcomes versus trichometric measurements in PRP therapy. Arch Dermatol Res. 2025

Chamberlain AJ, Dawber RP. Methods of evaluating hair growth. Australas J Dermatol. 2003

Saraogi PP, Dhurat RS. Automated digital image analysis (TrichoScan) for human hair growth analysis: ease versus errors. Int J Trichology. 2010

Majd A et al. A novel automated approach to rapid and precise in vivo measurement of hair morphometrics using a smartphone. Skin Res Technol. 2021

Pyo S et al. Automated measurement of hair density and projected apparent width via deep learning-based phototrichogram analysis. Int J Cosmet Sci. 2026

Huang J et al. Machine learning applications in androgenetic alopecia: a systematic review. Skin Health Dis. 2026

Alanazi SK et al. Artificial intelligence in trichology: a systematic review. Digit Health. 2026

Xi H et al. Enhanced stratification of male pattern hair loss using AI through novel loss region ratio analysis. Sci Rep. 2025

Suh MJ et al. Automated early detection of androgenetic alopecia using deep learning on trichoscopic images from a Korean cohort. Ewha Med J. 2025

Zhang S et al. Automated quantification of shedding hairs from smartphone photographs using an AI tool for alopecia monitoring. Clin Cosmet Investig Dermatol. 2026

De la Mettrie R et al. Shape variability and classification of human hair: a worldwide approach. Hum Biol. 2007

Loussouarn G et al. Worldwide diversity of hair curliness: a new method of assessment. Int J Dermatol. 2007

Mkentane K et al. Geometric classification of scalp hair for valid drug testing, 6 more reliable than 8 hair curl groups. PLoS One. 2017

Lasisi T et al. Quantifying variation in human scalp hair fiber shape and pigmentation. Am J Phys Anthropol. 2016

Callender V et al. Classification of high curl pattern hair: a systematic review and clinical perspective. J Cosmet Dermatol. 2026

Wortmann FJ, Stapels M, Chandra L. Modeling the time-dependent water wave stability of human hair. J Cosmet Sci. 2010

Breakspear S, Noecker B, Popescu C. Hair relaxation after shaping: a kinetic approach. Int J Cosmet Sci. 2024

Gao T. Evaluation of hair humidity resistance/moisturization from hair elasticity. J Cosmet Sci. 2007

Skedung L et al. A Curly Q: Is Frizz a Matter of Friction? Perception. 2021

Daniels G et al. Artificial intelligence in hair research: a proof-of-concept study on evaluating hair assembly features. Int J Cosmet Sci. 2021

Harrar M et al. A novel multiview image analysis methodology for quantifying hair volume on a human head (curl type I to III). Int J Cosmet Sci. 2026

Muthunilavan N et al. Defining the perceptual threshold for hair volume. Int J Cosmet Sci. 2026

Cruz CF et al. Effect of a peptide in cosmetic formulations for hair volume control. Int J Cosmet Sci. 2017

Kaplan PD et al. The shine problem in hair: review of imaging methods and measures for luster. J Cosmet Sci. 2009

Keis K, Ramaprasad KR, Kamath YK. Effect of hair color on luster. J Cosmet Sci. 2004

Keis K, Ramaprasad KR, Kamath YK. Studies of light scattering from ethnic hair fibers. J Cosmet Sci. 2004

McMullen R, Jachowicz J. Optical properties of hair: effect of treatments on luster as quantified by image analysis. J Cosmet Sci. 2003

McMullen R, Jachowicz J. Optical properties of hair: detailed examination of specular reflection patterns in various hair types. J Cosmet Sci. 2004

Lefaudeux N et al. New luster formula for the characterization of hair tresses using polarization imaging. J Cosmet Sci. 2009

Lim JM et al. A study correlating between instrumental and consumers' subjective luster values in oriental hair tresses. J Cosmet Sci. 2006

Sinclair RD. Healthy hair: what is it? J Investig Dermatol Symp Proc. 2007

Rudnicka L et al. Trichoscopy in hair shaft disorders. Dermatol Clin. 2018

Rudnicka L et al. Hair shafts in trichoscopy: clues for diagnosis of hair and scalp diseases. Dermatol Clin. 2013

Talghini S, Ranjkesh M. Microscopic hair changes associated with hair coloring, hair waving and hair ironing in Iranian women. Pak J Biol Sci. 2013

Taylor D et al. The biomechanics of splitting hairs. 2024

Singal A, Daulatabad D, Grover C. Graying severity score: a useful tool for evaluation of premature canities. Indian Dermatol Online J. 2016

Rosenberg AM et al. Quantitative mapping of human hair greying and reversal in relation to life stress. eLife. 2021

Bacon RA, Mizoguchi H, Schwartz JR. Assessing therapeutic effectiveness of scalp treatments for dandruff and seborrheic dermatitis, part 1: the adherent scalp flaking score (ASFS). J Dermatolog Treat. 2014

Ward B et al. Nasal distortion in short-distance photographs: the selfie effect. JAMA Facial Plast Surg. 2018

Derakhshan A et al. Quantifying facial distortion in modern digital photography. Laryngoscope. 2024

Related