Skin

Skin pigmentation, surface morphology, and visible quality — formalizing the Fitzpatrick treatment that previously lived only as a column in the 14-field per-image analysis schema. Captures the dimensions that drive skin-related ethnic variation: Fitzpatrick skin type, undertone, pigmentation pattern, acral pigmentation differential, and visible aging. The Fitzpatrick scale (1988) is the canonical medical reference for skin classification across populations and remains the load-bearing dimension here.

Dimensions

• Fitzpatrick skin type

  photo-observable

  ordinal · fitzpatrick_1988

  Six-class ordinal classification of constitutive skin pigmentation, originally developed for sun-reactivity assessment but now the de-facto skin-tone vocabulary in dermatology and computer-vision fairness research.

  Fitzpatrick TB (1988). The validity and practicality of sun-reactive skin types I through VI. Archives of Dermatology, 124(6): 869-871.

  Valid values (11)

  • IType I — very fair— Always burns, never tans. Pale white skin, often with red or blonde hair, blue/green eyes, freckles. Common in Northern European populations of Celtic / Scandinavian descent.
  • IIType II — fair— Burns easily, tans minimally. Fair skin, light hair, light eyes. Common in Northern European populations.
  • IIIType III — medium— Sometimes burns, gradually tans. Light olive to medium skin tone. Common in Mediterranean European, Middle Eastern, and East Asian populations.
  • IVType IV — olive / light brown— Burns minimally, tans easily and well. Olive to light brown skin. Common in Mediterranean, Middle Eastern, Latin American, Southeast Asian populations.
  • VType V — brown— Rarely burns, tans deeply. Brown skin tone. Common in South Asian, North African, Middle Eastern, some Latin American populations.
  • VIType VI — deeply pigmented brown to black— Never burns, deeply pigmented at baseline. Common in Sub-Saharan African, African-diaspora, some South Asian, some Pacific Islander populations.
  • II-IIIRange II-III— Bridging type used when assessment indicates the subject sits between Types II and III. Acceptable per the original Fitzpatrick description.
  • III-IVRange III-IV— Bridging type used when assessment indicates the subject sits between Types III and IV.
  • IV-VRange IV-V— Bridging type used when assessment indicates the subject sits between Types IV and V.
  • V-VIRange V-VI— Bridging type used when assessment indicates the subject sits between Types V and VI.
  • unclearUnclear— Lighting, image quality, makeup, or other obscuration prevents reliable Fitzpatrick assessment.

• Skin undertone

  partly photo-observable

  categorical · undertone_qualitative

  Hue character of the skin independent of overall depth (Fitzpatrick). Reflects relative balance of hemoglobin (red), melanin (brown), and carotenoid (yellow) contributions to perceived color.

  Cosmetic-industry standard color theory; aligned with anatomic descriptors used in Hauben & Mahler (1983) and dermatological literature on hemoglobin vs melanin contributions to skin color.

  Valid values (5)

  • warmWarm (yellow / golden)— Yellow, peach, or golden cast. Veins on inner wrist appear greenish.
  • coolCool (pink / red / blue)— Pink, rosy, or bluish cast. Veins on inner wrist appear blue or purple.
  • neutralNeutral— Balanced; neither warm nor cool dominant.
  • oliveOlive— Greenish-yellow cast; common in Mediterranean, Middle Eastern, and some South / Southeast Asian populations. Often grouped with warm in cosmetic shorthand but anatomically distinct.
  • unclearUnclear— Lighting or color balance prevents reliable assessment.

• Surface texture

  partly photo-observable

  ordinal · surface_texture_qualitative

  Visible surface texture of the facial skin — smoothness or coarseness as observable from a high-resolution close-up.

  Aligned with dermatologic surface-texture descriptors and the Visual Roughness scales used in cosmeceutical research (e.g. Robinson 1997).

  Valid values (5)

  • very_smoothVery smooth— No visible pore structure; flawless surface. Most often a function of image quality / soft-focus / makeup rather than ground-truth skin state.
  • smoothSmooth— Fine, even surface; visible pores small and uniform.
  • normalNormal— Average surface texture; pores visible but not prominent.
  • texturedTextured— Visible pore prominence, scarring, or surface irregularity.
  • coarseCoarse— Pronounced surface roughness, visible pore enlargement, or significant surface irregularity.

• Pigmentation pattern

  photo-observable

  categorical · pigmentation_pattern_qualitative

  Distribution pattern of skin pigmentation independent of overall Fitzpatrick depth.

  Categorical descriptors aligned with dermatology-textbook treatment of constitutional and acquired pigmentation patterns (Bolognia, Jorizzo, Schaffer — Dermatology, 4th Edition, 2018).

  Valid values (8)

  • uniformUniform— Even pigmentation across visible skin; no obvious lighter or darker patches.
  • freckled_ephelidesFreckled (ephelides)— Visible ephelides — small, well-circumscribed, sun-induced macules. Concentrated on sun-exposed areas; common in Fitzpatrick I-III.
  • lentiginesLentigines— Larger, persistent macules independent of sun exposure. Common with photoaging; can also be constitutional.
  • mottledMottled— Irregular patches of varying pigmentation without a single named pattern.
  • melasma_centrofacialMelasma — centrofacial— Hyperpigmented patches across the central face (forehead, cheeks, upper lip, chin); common in Fitzpatrick III-V.
  • post_inflammatory_hyperpigmentationPost-inflammatory hyperpigmentation (PIH)— Darker patches following inflammation; common in Fitzpatrick IV-VI.
  • vitiliginousVitiligo / depigmented patches— Visible loss of pigmentation in well-circumscribed patches.
  • unclearUnclear— Image conditions prevent reliable assessment.

• Freckling density

  photo-observable

  ordinal · freckle_density_qualitative

  Density of visible freckles (ephelides) on the face. A separate dimension from pigmentation_pattern because freckling density varies continuously and is meaningful even within the freckled category.

  Aligned with the visual-grading approach used in MC1R polymorphism / freckle phenotype research (Bastiaens et al. 2001). No single canonical scale; descriptive bucket vocabulary.

  Valid values (4)

  • noneNone— No discernible freckles.
  • fewFew— Sparse freckles, scattered, easy to count individually.
  • moderateModerate— Clearly freckled but countable in clusters.
  • heavyHeavy— Dense freckling; individual freckles begin to merge into larger pigmented areas.

• Acral pigmentation differential

  rarely from photo

  categorical · acral_pigmentation_qualitative

  Difference between palmar/plantar (acral) skin pigmentation and overall body pigmentation. Strong constitutional dimension in higher-Fitzpatrick populations: palms and soles are typically substantially lighter than the rest of the body in Fitzpatrick V-VI individuals, and this differential is itself a phenotype dimension.

  Constitutional palmar / plantar pigmentation differential well-documented in dermatology textbooks; relevant ethnic-anthropology dimension. Bolognia, Jorizzo, Schaffer — Dermatology, 4th Edition, 2018.

  Valid values (4)

  • minimalMinimal differential— Palmar/plantar skin tone similar to overall body tone; typical in lower Fitzpatrick types where the differential is small in absolute terms.
  • moderateModerate differential— Palms and soles visibly lighter than body skin; differential ~1-2 Fitzpatrick steps.
  • markedMarked differential— Palms and soles substantially lighter than body skin; differential 2+ Fitzpatrick steps. Typical in Fitzpatrick V-VI populations.
  • not_assessableNot assessable— Hands and feet not visible in photograph.

• Visible photoaging

  partly photo-observable

  ordinal · glogau_1996

  Glogau classification of photoaging severity, ordinal I-IV. Captures the visible cumulative effect of UV exposure on skin morphology — fine lines, deeper rhytides, dyschromia, and skin laxity.

  Glogau RG (1996). Aesthetic and anatomic analysis of the aging skin. Seminars in Cutaneous Medicine and Surgery, 15(3): 134-138.

  Valid values (5)

  • none_glogau_0None (subclinical)— No visible photoaging; smooth, even-toned skin. Typical of younger subjects with limited UV exposure.
  • glogau_IGlogau I — early photoaging— Mild dyschromia; minimal wrinkles. No visible keratoses. Typical age range 20s-30s.
  • glogau_IIGlogau II — early-moderate photoaging— Wrinkles in motion only ('expression lines'). Some lentigines visible. Typical age range 30s-40s.
  • glogau_IIIGlogau III — advanced photoaging— Wrinkles at rest. Visible dyschromia and telangiectasias. Typical age range 50s-60s.
  • glogau_IVGlogau IV — severe photoaging— Wrinkles throughout; sallow color; possible actinic keratoses. Typical age range 60s-70s+.

• Baseline erythema (visible redness)

  partly photo-observable

  ordinal · erythema_qualitative

  Visible facial redness independent of inflammatory state. Captures both constitutional ruddiness (common in Northern European populations) and rosacea-spectrum erythema as a continuum.

  Aligned with the Clinician's Erythema Assessment (CEA) scale used in rosacea grading (Tan J et al. 2014, Journal of the American Academy of Dermatology, 71). Adapted to non-rosacea baseline assessment by treating CEA grades 0-1 as constitutional.

  Valid values (4)

  • absentAbsent— No visible erythema; even baseline tone.
  • mildMild— Subtle pink/red flush, particularly on cheeks; constitutional ruddiness.
  • moderateModerate— Clear visible erythema across cheeks, central face; possible early rosacea.
  • markedMarked— Pronounced redness with visible telangiectasias; rosacea-spectrum erythema.