The evidence base

What the score is
actually built on.

Every peer-reviewed paper behind Noova’s skin scoring — 129 of them, listed in full. For each one: what the study found, and what we took from it. Including the papers that told us what a camera can’t do.

129 papersevery DOI verifiedzero fabricated citations

Why we publish this

Plenty of beauty apps say they are “backed by science.” Almost none will show you the list. This is ours, in full, with links — so you can check it.

It also keeps us honest. A good part of this research describes the limits of what a phone camera can read: hydration and oil are measured properly with probes against the skin, not a lens; reading redness is confounded by skin tone. So Noova estimates the appearance of these signals, tells you when the lighting isn’t good enough to trust, and grades evenness rather than tone. The research is what tells us where to stop.

Hydration

14 papers

What skin hydration actually is, how it is properly measured — and therefore what a camera can and cannot claim.

  1. Stratum corneum hydration in healthy adult humans according to the skin area, age and sex: a systematic review and meta-analysis

    Samadi A, et al. · Journal of the European Academy of Dermatology and Venereology : JEADV, 2022 · view paper

    FoundPooled corneometer SC-hydration ranged from a low of 12.5 (95% CI 11.05-13.94) on the back to a high of 64.34 (95% CI 62.07-64.59) on the periorbital area.

    We tookSite-specific corneometer capacitance anchors (facial/periorbital run high) for calibrating a hydration signal; but corneometry is a contact capacitance probe, so a camera cannot directly read SC water content.

  2. Transepidermal water loss in healthy adults: a systematic review and meta-analysis update

    Akdeniz M, et al. · The British journal of dermatology, 2018 · view paper

    FoundTEWL reference values ranged from a low of 2.3 g/m2/h (95% CI 1.9-2.7) on breast skin to a high of 44.0 g/m2/h (95% CI 39.8-48.2) on the axilla across 212 studies.

    We tookSite-dependent TEWL reference ranges anchor the barrier-proxy interpretation of hydration; TEWL requires a probe measuring water flux, which a camera cannot capture.

  3. Devices measuring transepidermal water loss: A systematic review of measurement properties

    Klotz T, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2022 · view paper

    FoundA reliability review examined 22 TEWL measurement devices across 38 included studies, finding good reliability but highly variable measurement error.

    We tookEstablishes that even instrumented TEWL has highly variable measurement error across 22 devices, setting an upper bound on precision; a camera-based hydration estimate cannot match probe-grade barrier measurement.

  4. Effects of Oral Collagen for Skin Anti-Aging: A Systematic Review and Meta-Analysis

    Pu SY, et al. · Nutrients, 2023 · view paper

    FoundIn 26 RCTs (1721 patients), hydrolyzed collagen significantly improved skin hydration (overall effect Z = 4.94, p < 0.00001) and elasticity (Z = 4.49, p < 0.00001) versus placebo.

    We tookQuantifies oral-collagen hydration effect size for treatment-response context; measured via corneometry not imaging, so the hydration gain is not directly visible to a camera.

  5. Effects of Collagen Supplements on Skin Aging: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

    Myung SK, et al. · The American journal of medicine, 2025 · view paper

    FoundA meta-analysis of 23 RCTs with 1474 participants found collagen improved hydration overall, but the effect vanished in non-industry-funded and high-quality studies.

    We tookCautions that collagen hydration benefits are funding/quality-dependent; relevant when weighting supplement claims but not a corneometry/TEWL anchor and unrelated to camera-readable signals.

  6. Effects of hydrolyzed collagen supplementation on skin aging: a systematic review and meta-analysis

    de Miranda RB, et al. · International journal of dermatology, 2021 · view paper

    FoundAcross 19 studies (1,125 participants aged 20-70, 95% women), 90 days of hydrolyzed collagen improved skin hydration, elasticity, and wrinkles versus placebo.

    We tookSupports a ~90-day timescale for measurable hydration change from ingestibles; hydration was instrument-measured, reinforcing that a camera cannot directly grade SC water content.

  7. Exploring the Impact of Hydrolyzed Collagen Oral Supplementation on Skin Rejuvenation: A Systematic Review and Meta-Analysis

    Dewi DAR, et al. · Cureus, 2023 · view paper

    FoundA meta-analysis of 14 studies (967 participants) found a 12-week hydrolyzed collagen regimen substantially enhanced skin moisture and elasticity versus placebo.

    We tookAdds a 12-week efficacy datapoint for hydration interventions; moisture measured instrumentally, so gains are not directly observable in a photo.

  8. The impact of dupilumab on skin barrier function: A systematic review

    Montero-Vilchez T, et al. · Journal of the European Academy of Dermatology and Venereology : JEADV, 2023 · view paper

    FoundAcross 6 studies (233 participants), dupilumab decreased TEWL, and 33.6% (2/6) reported increased stratum corneum hydration on eczematous lesions.

    We tookShows SCH and TEWL as the standard non-invasive readouts of barrier repair; both need probes, underscoring that a camera cannot directly read SC hydration.

  9. Restoration of Skin Barrier Abnormalities with IL4/13 Inhibitors and Jak Inhibitors in Atopic Dermatitis: A Systematic Review

    Chatzigeorgiou I, et al. · Medicina (Kaunas, Lithuania), 2024 · view paper

    FoundA review of 12 studies (378 dupilumab and 38 JAK-inhibitor participants) found significant TEWL decreases and stratum corneum hydration increases on eczematous lesions.

    We tookReinforces TEWL/SCH as coupled barrier-and-hydration endpoints in treatment response; both are instrument-based, not camera-derived.

  10. Clinical Measurement of Transepidermal Water Loss

    Kundu D, et al. · Advances in wound care, 2026 · view paper

    FoundThe FDA defines complete wound closure as 100% re-epithelialization verified across two assessments at least 14 days apart, yet high post-healing TEWL still predicts recurrence.

    We tookIllustrates that visible surface healing (camera-observable) can diverge from functional barrier status, which only TEWL detects, mirroring the limit of imaging-based hydration scoring.

  11. Transepidermal water loss (TEWL): Environment and pollution-A systematic review

    Green M, et al. · Skin health and disease, 2022 · view paper

    FoundAcross 15 studies (11 on environment/season, 4 on pollution), all four pollution studies showed increased TEWL with particulate-matter or NO2 exposure.

    We tookFlags environmental confounders (pollution, season) that shift TEWL/barrier state; useful context for interpreting hydration variability, and a factor a static camera image cannot control for.

  12. The Skin Barrier and Moisturization: Function, Disruption, and Mechanisms of Repair

    Rajkumar J, et al. · Skin pharmacology and physiology, 2023 · view paper

    FoundThe review models the skin barrier as four interdependent layers (physical, chemical, microbiologic, immunologic), with moisturizers shown to improve transepidermal water loss.

    We tookMechanistic framing for how occlusives/humectants/emollients lower TEWL and raise SC hydration, supporting product-match logic; the underlying hydration change is measured by probe, not imaging.

  13. Facial skin ageing: Key concepts and overview of processes

    Zargaran D, et al. · International journal of cosmetic science, 2022 · view paper

    FoundA review of 65 papers found combined intrinsic and extrinsic aging pathways impair skin barrier function alongside wrinkles, laxity, and thinning.

    We tookContext that aging degrades barrier/hydration function; relevant to age-adjusting hydration expectations but provides no corneometry/TEWL number and no camera-readable metric.

  14. Clinical Efficacy of a Salicylic Acid-Containing Gel on Acne Management and Skin Barrier Function: A 21-Day Prospective Study

    Liu Y, et al. · Journal of cosmetic dermatology, 2025 · view paper

    FoundIn 42 acne participants using a salicylic-acid gel for 21 days, Sebumeter SM815 sebum levels decreased by 23.65% while hydration rose 40.5%.

    We tookAnchors magnitude of achievable sebum change (~24% over 3 weeks) measured on Sebumeter SM815 in oily/combination skin; calibrates what an 'improving oil-balance' trajectory looks like.

Clarity

21 papers

The clinical scales used to grade breakouts, and how reliably photo-based systems reproduce a dermatologist’s grade.

  1. Prevalence and risk factors of acne scars in patients with acne vulgaris

    Liu L, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2023 · view paper

    FoundThe pooled prevalence of acne scars among acne patients was 47% (95% CI 38-56%), with severe acne carrying an odds ratio of 5.51 for scarring.

    We tookAnchors how often blemishes progress to permanent scars; clarity scoring should weight active-lesion severity as a downstream scarring predictor.

  2. Topical Retinoids in Acne Vulgaris: A Systematic Review

    Kolli SS, et al. · American journal of clinical dermatology, 2019 · view paper

    FoundA topical retinoid plus an oral antibiotic reduced acne lesion counts by 64-78.9% versus 41-56.8% for vehicle (p < 0.001).

    We tookGives realistic lesion-count reduction ranges over a treatment course, calibrating how much clarity change a scan should expect between visits.

  3. Management of Acne Vulgaris: A Review

    Eichenfield DZ, et al. · JAMA, 2021 · view paper

    FoundIn a 207-patient RCT, tretinoin 0.025% gel monotherapy reduced acne lesion counts by 63% versus baseline at 12 weeks.

    We tookProvides a concrete 12-week lesion-clearance benchmark (63%) that clarity trend deltas can be calibrated against.

  4. Systematic review of the epidemiology of acne vulgaris

    Heng AHS, et al. · Scientific reports, 2020 · view paper

    FoundOverweight/obese BMI (pooled OR 2.36, 95% CI 1.97-2.83) and parental family history (pooled OR 2.91, 95% CI 2.58-3.28) were strongly associated with acne presentation/severity.

    We tookQuantifies non-visual risk factors (BMI, family history) that clarity scoring cannot read from a photo but that contextualize baseline blemish likelihood.

  5. Epidemiology of acne vulgaris

    Bhate K, et al. · The British journal of dermatology, 2013 · view paper

    FoundAcne persists into the 20s and 30s in around 64% and 43% of individuals respectively, and moderate-to-severe acne affects around 20% of young people.

    We tookShows clarity is an ongoing adult concern, not adolescent-only, supporting age-agnostic clarity grading expectations.

  6. Safety and Effectiveness of Oral Nutraceuticals for Treating Acne: A Systematic Review

    Shields A, et al. · JAMA dermatology, 2023 · view paper

    FoundOf 2582 abstracts screened, 42 RCTs (3346 participants) met inclusion, with fair/good-quality studies showing decreased lesion counts or improved IGA for vitamins B5/D, green tea, probiotics and omega-3.

    We tookEvidence base for oral adjuncts tied to lesion-count/IGA outcomes, informing which non-topical factors could plausibly shift a clarity score.

  7. Guidelines of care for the management of acne vulgaris

    Reynolds RV, et al. · Journal of the American Academy of Dermatology, 2024 · view paper

    FoundThe 2024 AAD acne guideline issued 18 evidence-based recommendations and 5 good practice statements, with strong recommendations for benzoyl peroxide, topical retinoids, topical antibiotics and oral doxycycline.

    We tookAuthoritative treatment framework mapping severity to interventions, useful for grounding clarity-based product/routine recommendations.

  8. "Acne vulgaris: key insights, treatment, and future prospects"

    Bhatt J, et al. · Molecular diversity, 2026 · view paper

    FoundAcne vulgaris is the 8th most treated skin condition and affects over 80% of teenagers globally.

    We tookEstablishes near-universal adolescent prevalence, so a clear reading in that group is meaningfully above baseline expectation.

  9. Acne-induced Post-inflammatory Hyperpigmentation: From Grading to Treatment

    Auffret N, et al. · Acta dermato-venereologica, 2025 · view paper

    FoundAcne-induced post-inflammatory hyperpigmentation is mainly observed in patients with Fitzpatrick Phototypes IV to VI.

    We tookCamera/skin-tone caveat: residual pigment from cleared blemishes concentrates in darker phototypes, so clarity scoring must separate active lesions from PIH by skin tone.

  10. Diet and acne: A systematic review

    Meixiong J, et al. · JAAD international, 2022 · view paper

    FoundFrom 410 articles screened, 34 met inclusion; high glycemic index and glycemic load were positively associated with acne severity, supported by RCTs.

    We tookDiet is a modifiable, non-visual driver of blemish severity that clarity scoring cannot read directly but can flag as a contributing factor.

  11. Adolescent acne vulgaris: current and emerging treatments

    Layton AM, et al. · The Lancet. Child & adolescent health, 2023 · view paper

    FoundAcne vulgaris has a peak prevalence between age 15 and 20 years across all ethnicities and races.

    We tookDefines the age window where lower clarity is statistically expected, useful for age-normalized clarity interpretation.

  12. Low-dose oral isotretinoin for the treatment of adult patients with mild-to-moderate acne vulgaris: Systematic review and meta-analysis

    Al Muqarrab F, et al. · Dermatologic therapy, 2022 · view paper

    FoundConventional isotretinoin dosing is 0.5-1.0 mg/kg per day to a cumulative dose of 120-150 mg/kg, and conventional dosing improved prolonged-remission odds over low doses.

    We tookSeverity-treatment threshold context: isotretinoin marks the severe end of the clarity spectrum where sustained clearance is the goal.

  13. Update to acne vulgaris treatment for Canadian practice

    Keow S, et al. · Canadian family physician Medecin de famille canadien, 2025 · view paper

    FoundThe Canadian update applied the 2024 AAD acne guidelines, reviewing approved treatments for patients aged 9 and older with GRADE-rated evidence.

    We tookReinforces severity-tiered treatment mapping (topical for mild, oral for moderate-severe) that clarity-graded routine suggestions can follow.

  14. Development and validation of an artificial intelligence-powered acne grading system incorporating lesion identification

    Li J, et al. · Frontiers in medicine, 2023 · view paper

    FoundAn AI acne severity grader reached a linear weighted kappa of 0.652 against physician labels, rising to 0.737 and 0.696 once lesion-identification results were integrated with fixed and learnable weights.

    We tookConcrete reliability anchor: automated image-based acne severity grading tops out around weighted kappa 0.65-0.74 vs physicians, and explicit lesion detection (precision 0.507/recall 0.775) meaningfully lifts agreement, calibrating what noova's camera-driven clarity score can credibly claim.

  15. Artificial Intelligence in the Assessment and Grading of Acne Vulgaris: A Systematic Review

    Traini DO, et al. · Journal of personalized medicine, 2025 · view paper

    FoundAcross 29 included studies, AI acne diagnosis/lesion-counting/severity-grading models reached accuracy as high as 97.6% in controlled settings.

    We tookCeiling reference for image-based clarity scoring: peak 97.6% accuracy is controlled-setting only, and the review flags dataset/ethnic-bias and external-validation gaps that bound how far noova can trust a camera-only acne clarity read across diverse users.

  16. Inter-rater variability and consistency within four acne grading systems recommended in China, USA, and Europe

    Yu R, et al. · Journal of cosmetic dermatology, 2022 · view paper

    FoundFour acne grading systems (AGS, IIGS, IGA, GEA) applied by four investigators across 1107 acne patients showed excellent reliability and were significantly correlated with each other (p ≤ 0.01).

    We tookSupports interchangeability of major acne grading anchors (including US IGA and European GEA): different validated severity scales track together (p ≤ 0.01), so noova can map its clarity output onto IGA/GEA-style bands without conflicting ground truth.

  17. Comparisons of Four Acne Grading Systems Recommended in China, Korea, and Japan

    Yu R, et al. · Clinical, cosmetic and investigational dermatology, 2023 · view paper

    FoundAcross 1107 acne patients rated by five investigators, four grading systems (AGS, IIGS, KAGS, JAGS) showed strong correlations with one another (P ≤ 0.01).

    We tookReinforces cross-scale consistency for acne severity anchors used in China/Korea/Japan; strong inter-scale correlation (P ≤ 0.01) and reported strong intra-grader reliability mean a well-defined lesion-severity grade is reproducible enough to serve as a clarity training target.

  18. Acne Vulgaris and adherence to the mediterranean diet among university students: a case‒control study

    Taha S, et al. · Journal of health, population, and nutrition, 2024 · view paper

    FoundIn a GAGS-graded case-control study, family history of acne differed between cases and controls (OR = 2.01, CI 1.17-3.44) while Mediterranean-diet adherence did not (OR = 0.86, CI 0.46-1.60).

    We tookUses GAGS as the operational acne-severity instrument and quantifies non-image drivers: a camera reads lesion severity but cannot capture family history (OR 2.01), marking a clarity-score boundary where noova should defer to self-reported risk factors.

  19. Comparative Efficacy of Pharmacological Treatments for Acne Vulgaris: A Network Meta-Analysis of 221 Randomized Controlled Trials

    Huang CY, et al. · Annals of family medicine, 2023 · view paper

    FoundIn a network meta-analysis of 221 RCTs, oral isotretinoin produced the largest mean percentage reduction in total acne lesions (MD = 48.41), followed by topical retinoid + BPO triple therapy (MD = 38.15).

    We tookAnchors high %-match for retinoid-containing acne products; topical retinoid + BPO triple therapy (~38% lesion reduction) ranks near the top of 37 interventions.

  20. Topical benzoyl peroxide for acne

    Yang Z, et al. · The Cochrane database of systematic reviews, 2020 · view paper

    FoundTopical benzoyl peroxide improved participant-reported acne versus placebo with RR 1.27 (95% CI 1.12-1.45) over 10-12 weeks.

    We tookGrades BPO acne products; RR 1.27 improvement (little difference vs adapalene/clindamycin) supports a moderate %-match.

  21. Efficacy and Tolerability of a Novel Cosmetic and Over-the-Counter Facial Acne Regimen Versus a Prescription Treatment

    Huang P, et al. · Journal of cosmetic dermatology, 2025 · view paper

    FoundAn OTC salicylic acid acne regimen significantly improved Investigator's Global Assessment of acne severity from 48h through week 12 (p≤0.001), comparable to a prescription adapalene/BPO regimen.

    We tookSupports salicylic acid (BHA) OTC %-match; matched prescription adapalene/BPO on acne severity (p≤0.001) with better tolerance.

Texture

18 papers

Validated roughness and line-grading scales, and how a photograph compares against instrument measurement.

  1. Development and Validation of a Photonumeric Scale for Evaluation of Facial Skin Texture

    Donofrio L, et al. · Dermatologic surgery : official publication for American Society for Dermatologic Surgery [et al.], 2016 · view paper

    FoundIn a 290-subject validation of the 5-point Allergan Skin Roughness Scale, intrarater agreement across the 2 sessions was almost perfect at weighted kappa = 0.83.

    We tookAnchors a validated 5-grade midface roughness scale (kappa 0.83) noova can map camera-visible roughness bins to; a >=1-point change is clinically meaningful.

  2. Development and Validation of the Cheek Smoothness Scale: A Photonumeric Assessment Tool for Cheek Skin Texture

    Han HS, et al. · Journal of cosmetic dermatology, 2026 · view paper

    FoundThe Cheek Smoothness Scale is a five-grade photonumeric scale whose reliability was assessed by five board-certified dermatologists on single-angle (45 degrees) and multi-angle images, with grades tracking rising PRIMOS Ra values.

    We tookValidates a 5-grade smoothness scale against objective PRIMOS Ra and confirms a single 45-degree oblique photo suffices, supporting noova's single-angle camera texture grading.

  3. Skin measurement devices to assess skin quality: A systematic review on reliability and validity

    Langeveld M, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2022 · view paper

    FoundAcross the systematic review, skin texture was assessed in only two studies (72 subjects, three devices), with PRIMOS the sole device showing excellent intra- and interobserver reliability.

    We tookFlags a thin evidence base (2 studies / 72 subjects) for objective texture devices; PRIMOS is the reference instrument, but none reached construct validity, tempering how hard noova anchors absolute roughness.

  4. Collagen study advances for photoaging skin

    Liu H, et al. · Photodermatology, photoimmunology & photomedicine, 2024 · view paper

    FoundType I collagen makes up 80%-90% of skin collagen, type III 8%-12%, and type V 5%, and this matrix fragments and degrades with photoaging.

    We tookExplains the structural driver of surface texture loss (collagen fragmentation); camera reads the resulting roughness/wrinkles, not collagen composition itself.

  5. Cosmeceuticals in photoaging: A review

    Chan LKW, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2024 · view paper

    FoundIn this systematic review, cosmeceutical evidence for photodamaged skin ranged from Level I to Level II, with peptides strongest at Level Ib.

    We tookGrades strength of topical evidence for reversing textural photodamage; informs how confidently noova can recommend texture-targeting cosmeceuticals (peptides best-supported).

  6. Use of platelet rich plasma for skin rejuvenation

    Phoebe LKW, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2024 · view paper

    FoundPRP treatment produced significant enhancements in facial parameters including pore size, texture, and wrinkles after one to three sessions.

    We tookSupports pore-size and texture as responsive, camera-readable outcomes; the 1-3 session dose frames expected timelines for texture-improvement nudges.

  7. Sunscreens and Photoaging: A Review of Current Literature

    Guan LL, et al. · American journal of clinical dermatology, 2021 · view paper

    FoundThere is a paucity of US FDA-approved filters protecting against long UVA (>370 nm) and none against visible light, both implicated in photoaging.

    We tookCoverage gap above 370 nm means sunscreen may under-protect against texture/wrinkle photoaging; relevant to product-match guidance, not directly camera-readable.

  8. Development and Validation of Photo-Numerical Scales for Facial Wrinkles on Chinese Population

    Zhao X, et al. · Journal of cosmetic dermatology, 2025 · view paper

    FoundPhoto-numeric wrinkle grading scores correlated with Primos 3D metrics at r = 0.79 for wrinkle volume and r = 0.82 for wrinkle depth, with intra-grader ICC > 0.9 and inter-grader ICC > 0.8.

    We tookDirectly ties a camera-readable photonumeric wrinkle scale to objective 3D depth/volume roughness (r~0.8), anchoring how a phone image maps to true wrinkle geometry.

  9. Reliability and validity of two in vivo measurements for skin surface topography in aged adults

    Trojahn C, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2015 · view paper

    FoundIn vivo skin-roughness reliability spanned ICC 0.50-0.95 for Visioscan and 0.01-1.00 for PRIMOS, with Rmax, Rz, and Ra recommended as the most reliable topography parameters.

    We tookShows lab profilometry roughness (Ra/Rz) reliability is parameter-dependent and can be poor; a consumer camera cannot resolve these micro-roughness metrics, so texture must be inferred, not measured.

  10. The wrinkle severity rating scale: a validation study

    Day DJ, et al. · American journal of clinical dermatology, 2004 · view paper

    FoundThe Wrinkle Severity Rating Scale showed weighted kappa of 0.77 (left) and 0.81 (right) for intra-observer nasolabial-fold grading, with test-retest agreement of 68.7% and 72.7%.

    We tookValidated 5-grade wrinkle scale with substantial kappa reliability; a reference anchor for defining discrete fine-line/fold severity buckets in noova's texture scoring.

  11. Clinical evaluation of skin aging: a systematic review

    Wang YC, et al. · Archives of gerontology and geriatrics, 2025 · view paper

    FoundA systematic review of skin-aging scales included 42 publications and 47 scales: 20 for lines/wrinkles, 12 for sagging/volume, 6 for texture and quality, and 9 for global assessment.

    We tookMaps the validated-scale landscape and shows texture/quality has only 6 scales vs 20 for wrinkles, flagging where camera-readable, validated texture anchors are scarce.

  12. Development and Validation of a Photonumeric Scale for Evaluation of Lip Fullness

    Bloom J, et al. · Journal of drugs in dermatology : JDD, 2023 · view paper

    FoundThe Merz Lip Fullness Assessment Scale reached near-perfect intrarater agreement with median weighted kappa of 0.911 (upper lip) and 0.930 (lower lip), all kappas >= 0.6.

    We tookDemonstrates the kappa reliability ceiling (~0.9) achievable for a well-defined 5-point photonumeric facial scale; a methodological benchmark for noova's texture rating buckets.

  13. Facial Laxity Rating Scale Validation Study

    Leal Silva HG, et al. · Dermatologic surgery : official publication for American Society for Dermatologic Surgery [et al.], 2016 · view paper

    FoundThe Facial Laxity Rating scale had combined weighted-kappa inter-rater agreement of 0.77 (95% CI 0.76-0.78) across 188 ratings, with 92.53% agreement to the truth standard.

    We tookKappa ~0.77 for a trained-rater facial surface scale sets a realistic inter-rater reliability target that an automated texture/laxity classifier should aim to match.

  14. Clinical Trial of Salmon Nasal Cartilage-Derived Proteoglycans on Human Facial Antiaging: A Randomized, Double-Blind, Placebo-Controlled Study

    Bai XD, et al. · Journal of cosmetic dermatology, 2025 · view paper

    FoundIn a 56-day RCT (n=66), oral 20 mg salmon proteoglycan produced significant reductions in skin roughness and wrinkles (p < 0.001) versus placebo.

    We tookPlacebo-controlled evidence that measured roughness is a sensitive, movable texture endpoint over 8 weeks, supporting roughness as a trackable texture signal in noova.

  15. Efficacy of combined treatment with human adipose tissue stem cell-derived exosome-containing solution and microneedling for facial skin aging: A 12-week prospective, randomized, split-face study

    Park GH, et al. · Journal of cosmetic dermatology, 2023 · view paper

    FoundIn a split-face RCT (n=28), the exosome-plus-microneedling side scored significantly higher on the Global Aesthetic Improvement Scale than control (p = 0.005), with PRIMOS-confirmed wrinkle gains.

    We tookUses PRIMOS to quantify wrinkle/texture change in a controlled split-face design; reinforces that instrument-grade wrinkle metrics are the ground truth camera scoring approximates.

  16. Low-molecular-weight collagen peptides supplement promotes a healthy skin: A randomized, double-blinded, placebo-controlled study

    Seong SH, et al. · Journal of cosmetic dermatology, 2024 · view paper

    FoundIn a placebo-controlled RCT of 100 participants, collagen peptides significantly improved average skin roughness and peak-to-valley wrinkle metrics versus placebo by Week 12.

    We tookAdds RCT support that average roughness and peak-to-valley wrinkle depth are responsive texture endpoints, useful as target metrics behind noova's texture score.

  17. "The efficacy and adverse effects of treatment options for facial pores: A review article"

    Parvar SY, et al. · Journal of cosmetic dermatology, 2023 · view paper

    FoundA review of 19 clinical trials pooled 591 cases (83.7% female) aged 18-80 years evaluating treatments that reduce facial pores driven by sebum.

    We tookHigh sebum excretion drives visible pores; supports treating enlarged pores as a camera-readable proxy for oil and calibrating focus by age (sebum control young vs rejuvenation older).

  18. Facial Pores: Definition, Causes, and Treatment Options

    Lee SJ, et al. · Dermatologic surgery : official publication for American Society for Dermatologic Surgery [et al.], 2016 · view paper

    FoundThere are 3 major clinical causes of enlarged facial pores: high sebum excretion, decreased peri-pore elasticity, and increased hair follicle volume.

    We tookHigh sebum excretion is one of only three enlarged-pore drivers, so a camera reading enlarged pores partially indexes oil but must be disambiguated from elasticity/follicle causes.

Oil balance

14 papers

How oil actually behaves by age, zone and skin type — and why visible shine is not the same as a sebum measurement.

  1. Current Insights on Lipidomics in Dermatology: A Systematic Review

    Sanabria-de la Torre R, et al. · The Journal of investigative dermatology, 2025 · view paper

    FoundA PRISMA systematic review screened 1013 references and included 48 studies covering 2651 participants with a mean age of 34.13 ± 16.28 years.

    We tookEstablishes sebutape/skin-surface lipidomics as the dominant in-vivo sebum sampling method and gives a pooled evidence base + age anchor for oil-signal literature.

  2. Skin care and rejuvenation by cosmeceutical facial mask

    Nilforoushzadeh MA, et al. · Journal of cosmetic dermatology, 2018 · view paper

    FoundFacial masks used to remove sebum fall into four categories: sheet, peel-off, rinse-off, and hydrogel masks.

    We tookWeak/categorical: informs the product-side of oil balance (masks physically remove surface sebum) rather than a measurable sebum value.

  3. [The sebaceous gland]

    Zouboulis CC, et al. · Der Hautarzt; Zeitschrift fur Dermatologie, Venerologie, und verwandte Gebiete, 2010 · view paper

    FoundSebum excretion rises with adrenarche at about age 9 years and continues climbing to the adult level by age 17 years.

    We tookGives the age window (9-17) over which sebum output ramps to adult levels — direct anchor for age-calibrating oil-balance expectations rather than flagging teens as anomalous.

  4. Can sebum reduction predict acne outcome?

    Janiczek-Dolphin N, et al. · The British journal of dermatology, 2010 · view paper

    FoundPooled across treatments, a 30% to 50% reduction in sebum excretion was projected to achieve a 50% reduction in acne measures (slopes significant, P < 0.025).

    We tookQuantifies the sebum-to-acne dose relationship, justifying oil as a meaningful upstream signal: moderate sebum shifts (30-50%) map to large clinical acne change.

  5. Acne and sebaceous gland function

    Zouboulis CC, et al. · Clinics in dermatology, 2004 · view paper

    FoundSebum excretion undergoes a new rise at about age 9 years with adrenarche and reaches adult levels by age 17 years.

    We tookCorroborates the 9-17 year sebum ramp and ties elevated sebum excretion to adolescent acne — supports age-band priors in oil scoring.

  6. The effect of 2% niacinamide on facial sebum production

    Draelos ZD, et al. · Journal of cosmetic and laser therapy : official publication of the European Society for Laser Dermatology, 2006 · view paper

    FoundIn an RCT, topical 2% niacinamide significantly reduced casual sebum level (CSL) after 6 weeks in Caucasians, while sebum excretion rate (SER) was not significantly reduced.

    We tookShows CSL and SER are distinct measurable oil endpoints that respond differently by ethnicity/ingredient — the camera reads casual sebum level (surface shine), not secretion rate.

  7. High-Resolution Spatial Map of the Human Facial Sebaceous Gland Reveals Marker Genes and Decodes Sebocyte Differentiation

    Düz T, et al. · The Journal of investigative dermatology, 2026 · view paper

    FoundSpatial transcriptomics of the human facial sebaceous gland identified 4 distinct stages of sebocyte differentiation, each with unique gene signatures.

    We tookWeak/structural: documents human (not mouse) sebaceous-gland biology underlying sebum output; background support for oil, not a surface-measurable value.

  8. Biometric Evaluation of Baumann Skin Typing

    Delavar S, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2026 · view paper

    FoundIn 71 volunteers, the Baumann oily-vs-dry questionnaire score correlated only weakly (though significantly) with measured sebum (r = 0.299, p = 0.011), while hydration and TEWL did not.

    We tookCritical calibration point: self-report/questionnaire oily-vs-dry typing tracks objective sebum only weakly (r=0.30), so noova's oil signal needs objective (camera/sebumeter) grounding, not questionnaire alone.

  9. The Efficacy and Safety of Chitosan on Facial Skin Sebum

    Theerawattanawit C, et al. · Skin pharmacology and physiology, 2022 · view paper

    FoundChitosan nanoparticle gel lowered T-zone sebum below placebo at week 4 (p = 0.043), with no difference in the U-zone.

    We tookReinforces T-zone vs U-zone regional sebum separation for facial-region calibration; only the T-zone moved significantly, supporting region-weighted oil scoring.

  10. Development and validation of a prospective questionnaire for assessing oily sensitive skin

    Zhang J, et al. · International journal of cosmetic science, 2024 · view paper

    FoundIn 1297 subjects, ROC-derived oily-skin cut-offs of 22.5 and 31.5 distinguish moderately and severely oily skin, with sebum levels differing significantly across the three oily groups.

    We tookProvides validated oiliness severity thresholds (moderate vs severe) tied to significantly different measured sebum, useful as tiering anchors for a self-reported/camera-inferred oil scale.

  11. Facial skin characteristics and concerns in Indonesia: A cross-sectional observational study

    Du Y, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2022 · view paper

    FoundAcross 419 Indonesian females, facial sebum stayed similar up to age 50 then declined in the 50-60 age band.

    We tookAge-calibration anchor: sebum is roughly flat through the 20s-40s and drops only after 50, so age should shift expected-sebum baselines mainly at older ages.

  12. Sebum excretion in hidradenitis suppurativa

    Jemec GB, et al. · Dermatology (Basel, Switzerland), 1997 · view paper

    FoundSebum excretion and active-gland counts followed a strong regional gradient of face > axillae/genitofemoral fold (p < 0.0001) in both patients and controls.

    We tookQuantifies that the face is the highest-sebum region body-wide, justifying facial-region weighting and that hidradenitis (non-facial) does not elevate facial sebum.

  13. Effect of antihistamine as an adjuvant treatment of isotretinoin in acne: a randomized, controlled comparative study

    Lee HE, et al. · Journal of the European Academy of Dermatology and Venereology : JEADV, 2014 · view paper

    FoundAdding antihistamine to isotretinoin cut total acne lesions 45.6% vs 18.7% at week 12, alongside a significant reduction in measured sebum.

    We tookLinks measured sebum reduction to lesion clearance, supporting sebum as an upstream driver of acne severity in oil-balance scoring.

  14. Skin Anti-Aging and Moisturizing Effects of Low-Molecular-Weight Collagen Peptide Supplementation in Healthy Adults: A Randomized, Double-Blind, Placebo-Controlled Clinical Trial

    Lee E, et al. · Journal of microbiology and biotechnology, 2025 · view paper

    FoundIn 70 adults, oral collagen peptide at 1,650 mg/day for 8 weeks produced decreased sebum secretion and pore-tightening effects.

    We tookEvidence that sebum and pore size covary and both respond to an intervention, supporting a coupled pore/oil sub-signal in scoring.

Redness

14 papers

Redness grading, and the confound that matters most: melanin interferes with reading redness, so deeper skin needs wider uncertainty.

  1. Transcutaneous Auricular Vagus Nerve Stimulation Treatment for Erythematotelangiectatic Rosacea: A Randomized Clinical Trial

    Li J, et al. · JAMA dermatology, 2025 · view paper

    FoundIn an RCT, mean CEA score at 3 weeks was 1.56 (SD 0.84) with taVNS vs 2.47 (SD 0.81) with sham (mean difference -0.92; 95% CI -1.3 to -0.53; P<.001).

    We tookAnchors the CEA 0-4 grading scale numerically and shows a ~0.9-point CEA change is a clinically meaningful erythema shift for calibrating noova's redness deltas.

  2. Meta-Analysis of the Efficacy of Intense Pulsed Light and Pulsed-Dye Laser Therapy in the Management of Rosacea

    Zhai Q, et al. · Journal of cosmetic dermatology, 2024 · view paper

    FoundAcross 4 studies (141 participants), change in erythema index did not differ between IPL and PDL (SMD = -0.15, 95% CI: -0.55 to 0.26).

    We tookConfirms erythema index is a standard pooled outcome for rosacea severity, supporting index-style redness readouts, though camera-based indices differ from clinical devices.

  3. Incidence and prevalence of rosacea: a systematic review and meta-analysis

    Gether L, et al. · The British journal of dermatology, 2018 · view paper

    FoundPooled rosacea prevalence was 5.46% (95% CI 4.91-6.04) in the general population and 2.39% (95% CI 1.56-3.39) among dermatology outpatients.

    We tookBase-rate prior for how often facial erythema/rosacea appears in a general population, useful for setting expected distribution of redness scores.

  4. Treatment of ocular rosacea: a systematic review

    Avraham S, et al. · Journal der Deutschen Dermatologischen Gesellschaft = Journal of the German Society of Dermatology : JDDG, 2024 · view paper

    FoundIn 1,275 patients, topical antimicrobials and oral antibiotics achieved complete or partial response in 91% (82/90) and 89% (525/580) of patients respectively.

    We tookQuantifies treatment-response rates for rosacea, informing how much calming/erythema improvement noova should expect after intervention for its redness-trend messaging.

  5. The role of vitamins and nutrients in rosacea

    Algarin YA, et al. · Archives of dermatological research, 2024 · view paper

    FoundOral isotretinoin produced a 90% reduction in lesions, complete remission in 24% of patients, and marked improvement in 57% of patients.

    We tookEffect-size anchors for redness/lesion improvement magnitudes, useful for calibrating what constitutes large vs modest calming in noova's redness deltas.

  6. Rosacea management: A comprehensive review

    Sharma A, et al. · Journal of cosmetic dermatology, 2022 · view paper

    FoundRosacea affects up to 10% of the world population, especially fair-skinned individuals aged 35-50.

    We tookPrevalence ceiling and demographic (fair skin) note; relevant because camera erythema reads differ by baseline skin tone and age band.

  7. Signaling pathways and targeted therapy for rosacea

    Yang F, et al. · Frontiers in immunology, 2024 · view paper

    FoundRosacea has a reported global prevalence ranging from 1% to 20%.

    We tookWide prevalence range documents how much erythema base-rate varies by population, cautioning against a single fixed redness threshold across cohorts.

  8. Outcome measurement instruments for erythema associated with incontinence-associated dermatitis: Systematic review

    De Meyer D, et al. · Journal of advanced nursing, 2019 · view paper

    FoundOf 14 studies covering 10 erythema instruments for incontinence-associated dermatitis, the most-studied measurement properties were reliability (9 studies), measurement error (4 studies), and criterion validity (4 studies).

    We tookShows erythema measurement instruments are validated mainly on reliability, flagging that redness grading tools need explicit reliability/measurement-error evidence.

  9. Neurogenic Rosacea Treatment: A Literature Review

    Ivanic MG, et al. · Journal of drugs in dermatology : JDD, 2023 · view paper

    FoundA systematic search identified 6 articles with a total of 37 patients with suspected neurogenic rosacea.

    We tookSmall evidence base for the flushing/erythema-dominant neurogenic subtype, a reminder that some erythema presentations are under-characterized clinically.

  10. Treatment management for rosacea: current pharmacological and non-pharmacological options

    Volk K, et al. · Expert review of clinical pharmacology, 2025 · view paper

    FoundRosacea is classified into four subtypes: erythematotelangiectatic, papulopustular, phymatous, and ocular involvement.

    We tookDefines the four-subtype framework so a redness signal can be mapped to the erythematotelangiectatic subtype that a camera reads most directly.

  11. Reliability of Clinician Erythema Assessment grading scale

    Tan J, et al. · Journal of the American Academy of Dermatology, 2014 · view paper

    FoundThe Clinician's Erythema Assessment 5-point scale had an overall intraclass correlation coefficient of 0.601 (session 1) and 0.576 (session 2), with an overall weighted kappa of 0.692.

    We tookCEA is the clinical 5-point erythema anchor; ICC ~0.58-0.60 among trained dermatologists sets a realistic reliability ceiling that noova's redness score can be benchmarked against.

  12. Reliability of a Skin Diagnostic Device in Assessing Hydration and Erythema

    Huimin K, et al. · Advances in skin & wound care, 2017 · view paper

    FoundThe SD202 skin diagnostic device showed high internal consistency for consecutive erythema measures at a single venous leg ulcer periwound point (Cronbach alpha > .970).

    We tookInstrument-based erythema reads (SD202) reach very high repeatability (alpha > .97), a device benchmark for noova's redness read, though the abstract notes consistency dropped in predebridement measures.

  13. Evaluation of different factors influencing objective measurement of skin color by colorimetry

    Anbar TS, et al. · Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging (ISSI), 2019 · view paper

    FoundIncomplete device-skin contact significantly lowered the erythema index (P < 0.001) while high applied pressure significantly raised it (P < 0.001) on colorimetric measurement.

    We tookErythema-index reads are strongly confounded by device contact and pressure (P < 0.001); a camera-based redness read must control geometry/contact-equivalent artifacts and cannot assume EI is pressure-invariant.

  14. Validation of a neonatal skin color scale

    Maya-Enero S, et al. · European journal of pediatrics, 2020 · view paper

    FoundA neonatal skin-color scale validated against the Mexameter MX 18 achieved 83.2% inter-observer agreement, with only melanin (not erythema) differing significantly across color groups.

    We tookShows Mexameter melanin index cleanly separates tone categories independent of erythema, validating melanin-index as the anchor for tone (not redness) in evenness scoring.

Pigment

22 papers

Measuring evenness within a person’s own tone — never lightness — plus the limits of agreement between devices.

  1. Laser therapy in the treatment of melasma: a systematic review and meta-analysis

    Lai D, et al. · Lasers in medical science, 2022 · view paper

    FoundLow-fluence Q-switched 1,064-nm Nd:YAG laser reduced MASI with a weighted mean difference of -2.76 (95% CI: -3.53 to -1.99).

    We tookGives MASI effect-size anchors (WMD) for pigment-lesion change; MASI is a clinician grader, not a camera read, useful for calibrating the expected magnitude of a real evenness shift.

  2. Efficacy and Safety of Tranexamic Acid in Melasma: A Meta-analysis and Systematic Review

    Kim HJ, et al. · Acta dermato-venereologica, 2017 · view paper

    FoundTranexamic acid alone lowered MASI by 1.60 (95% CI 1.20-2.00; p<0.001).

    We tookQuantifies a ~1.6-point MASI drop as a clinically meaningful pigment change, anchoring what counts as a real evenness improvement versus measurement noise.

  3. Comparison of the Efficacy of Melasma Treatments: A Network Meta-Analysis of Randomized Controlled Trials

    Liu Y, et al. · Frontiers in medicine, 2021 · view paper

    FoundAcross melasma therapies, side-effect rates ranged from 10.1% (tretinoin) to 52.3% (microneedles).

    We tookMASI-based ranking plus side-effect percentages across 14 therapies gives a comparative pigment-treatment reference, with MASI as the common evenness outcome metric.

  4. The clinical effect of glutathione on skin color and other related skin conditions: A systematic review

    Dilokthornsakul W, et al. · Journal of cosmetic dermatology, 2019 · view paper

    FoundOral glutathione 500 mg/day and topical 2.0% oxidized glutathione brightened sun-exposed skin as measured by melanin index.

    We tookDirectly validates melanin index (Mexameter-style) as a measurable brightening signal, and that change appears in sun-exposed (uneven) areas, reinforcing within-tone evenness rather than melanin-depth framing.

  5. Chemical Peels for Melasma: A Systematic Review

    Sarkar R, et al. · Dermatologic surgery : official publication for American Society for Dermatologic Surgery [et al.], 2024 · view paper

    FoundA review of 24 studies (15 RCTs) with a total sample of 1,075 patients quantified chemical-peel efficacy in melasma using MASI.

    We tookConfirms MASI as the standard pigment-grading outcome across a 1,075-patient evidence base, supporting MASI as the reference scale for melasma evenness.

  6. Topical treatment for postinflammatory hyperpigmentation: a systematic review

    Tan MG, et al. · The Journal of dermatological treatment, 2022 · view paper

    FoundA PIH systematic review included 47 of 1,224 screened studies covering 1,853 subjects.

    We tookAnchors the post-inflammatory hyperpigmentation evidence base and notes inflammation is subtle in darker phototypes, a caution that cameras may miss early uneven pigment in deep tones.

  7. Topical and Systemic Therapies in Melasma: A Systematic Review

    Sarkar R, et al. · Indian dermatology online journal, 2023 · view paper

    FoundA PRISMA melasma review identified 174 randomized/controlled trials grading hydroquinone, triple-combination cream, sunscreens, kojic and azelaic acid as grade A.

    We tookLarge evidence map of pigment therapies with grade-A recommendations, supporting an evidence-backed catalogue of evenness treatments to surface.

  8. Characteristics and management of Asian skin

    Chan IL, et al. · International journal of dermatology, 2019 · view paper

    FoundAsian skin spans Fitzpatrick types III to IV (Chinese, Japanese) to IV and V (Indian, Pakistani).

    We tookFitzpatrick III-V reference mapping across subgroups supports fair within-tone scoring across deep phototypes; the paper also notes melanocyte number is similar across types, reinforcing never penalizing melanin depth.

  9. Clinical Validation of Imaging Biomarkers in Mycosis Fungoides

    Wind SS, et al. · Experimental dermatology, 2026 · view paper

    FoundMSI individual typology angle (ITA) and companion imaging biomarkers showed strong test-retest reliability with CV < 10% and ICC > 0.84 while discriminating lesional from non-lesional skin at p ≤ 0.001.

    We tookAnchors ITA as a repeatable, camera-readable pigmentation biomarker (ICC > 0.84, CV < 10%), supporting ITA-based within-tone evenness scoring in noova.

  10. Clinical Evaluation of Thiamidol-Containing Formulations for the Visual Management of Facial Hyperpigmentation

    Desai SR, et al. · Journal of drugs in dermatology : JDD, 2025 · view paper

    FoundIn 90 subjects (n=43 serum; n=47 regimen), Thiamidol formulations produced significant increases in L* and ITA° values as early as week 2, with continued improvement through week 12.

    We tookConfirms ITA° (with L*) as a sensitive colorimetric readout of hyperpigmentation change over weeks, supporting ITA as a longitudinal evenness signal noova can track.

  11. The efficacy and safety of microneedling with topical tranexamic acid for melasma treatment: A systematic review and meta-analysis

    Feng X, et al. · Journal of cosmetic dermatology, 2024 · view paper

    FoundPooled microneedling-plus-topical-TXA reduced MASI/mMASI/hemiMASI vs baseline with MD = 1.85 (4 wk), 3.28 (8 wk), and 4.73 (12 wk).

    We tookProvides MASI change magnitudes (MD ~1.8-4.7) that calibrate what a clinically meaningful pigment-evenness improvement looks like on the MASI scale.

  12. Efficacy and safety of laser-related therapy for melasma: A systematic review and network meta-analysis

    Ma W, et al. · Journal of cosmetic dermatology, 2023 · view paper

    FoundNetwork meta-analysis of 39 studies (1394 participants) found QSND+topical medications superior to QSND alone by MD = -4.21 (-6.80, -1.63) on MASI.

    We tookQuantifies between-treatment MASI differences (MD ~-3.5 to -4.2), anchoring the resolution at which MASI distinguishes pigment outcomes.

  13. Application of PRP in Chloasma: A Meta-Analysis and Systematic Review

    Deng T, et al. · Computational intelligence and neuroscience, 2022 · view paper

    FoundIntradermal PRP for chloasma significantly decreased MASI (mean -6.71, 95% CI -8.99 to -4.33) and mMASI (mean -2.94, 95% CI -4.81 to -1.07).

    We tookGives paired MASI (-6.71) vs mMASI (-2.94) effect sizes, useful for mapping raw vs modified MASI scale magnitudes in pigment-evenness scoring.

  14. Efficacy of Botulinum Toxin A for the Management of Melasma: A Split-Face, Randomized Control Study

    Chaijaras S, et al. · Journal of cosmetic dermatology, 2025 · view paper

    FoundIn a split-face RCT, week-24 MASIm was 14.79 (21.03% worse) on monotherapy vs 9.14 (36.97% improvement, p = 0.0003) on the BoNT-A combination side.

    We tookSplit-face MASIm plus colorimeter melanin index gives paired within-subject percent-change anchors (36.97% vs -21.03%) for pigment scoring calibration.

  15. Tranexamic acid as a therapeutic option for melasma management: meta-analysis and systematic review of randomized controlled trials

    Calacattawi R, et al. · The Journal of dermatological treatment, 2024 · view paper

    FoundA meta-analysis of 22 studies with 1280 patients found TXA significantly reduced MASI, mMASI, MI, and hemi-MASI scores.

    We tookLarge pooled evidence base (1280 patients) confirming MASI/mMASI/MI as convergent pigment-severity endpoints noova can lean on.

  16. Comparative Efficacy and Safety of Injectable Tranexamic Acid Combination Therapies for Melasma: A Network Meta-analysis of Randomized Controlled Trials

    Nukaly H, et al. · Aesthetic surgery journal, 2025 · view paper

    FoundA network meta-analysis of 9 RCTs (358 participants) found TXA + 4% hydroquinone gave the largest MASI improvement of the compared interventions.

    We tookMASI-scored comparative evidence (358 participants) reinforcing MASI as the standard pigment-severity outcome for evenness benchmarking.

  17. Glutathione as an oral whitening agent: a randomized, double-blind, placebo-controlled study

    Arjinpathana N, et al. · The Journal of dermatological treatment, 2012 · view paper

    FoundOral glutathione 500 mg/day for 4 weeks lowered melanin index significantly more than placebo at two of six sites (p = 0.021 and 0.036).

    We tookDemonstrates melanin index sensitivity to modest, site-specific pigment change, reminding that melanin depth shifts are measurable but should never be penalized as unevenness.

  18. Microneedling as an adjuvant to topical therapies for melasma: A systematic review and meta-analysis

    Bailey AJM, et al. · Journal of the American Academy of Dermatology, 2022 · view paper

    FoundAdding microneedling to topical therapies (including vitamin C) improved melasma severity with a large effect (SMD >0.8) beyond 8 weeks.

    We tookInforms vitamin C/topical melasma %-match; microneedling-assisted topicals reach a large effect (SMD>0.8) on pigmentation by 12 weeks.

  19. Comparative Efficacy of 35 % Glycolic Acid Peel versus 4 % Retinol Peel in Melasma

    Rini T, et al. · Dermatologic surgery : official publication for American Society for Dermatologic Surgery [et al.], 2026 · view paper

    FoundIn melasma, mild self-limiting adverse effects were more common with 4% retinol peel (13.3%) than 35% glycolic acid peel (3.3%), with both peels comparably effective.

    We tookGrades AHA (glycolic) vs retinol peel %-match for melasma; comparable efficacy with glycolic better tolerated (3.3% vs 13.3% AEs).

  20. Evaluation of the Efficacy of a Serum Containing Niacinamide, Tranexamic Acid, Vitamin C, and Hydroxy Acid Compared to 4% Hydroquinone in the Management of Melasma

    Rocio J, et al. · Journal of cosmetic dermatology, 2025 · view paper

    FoundA niacinamide/tranexamic-acid/vitamin C serum significantly reduced melasma pigmentation (p<0.001) after 3 months, with efficacy similar to 4% hydroquinone.

    We tookStrong niacinamide + vitamin C %-match for pigment: a multi-active serum matched 4% hydroquinone (p<0.001) with better local tolerance.

  21. Glutathione as a skin-lightening agent and in melasma: a systematic review

    Sarkar R, et al. · International journal of dermatology, 2025 · view paper

    FoundFive RCTs plus one open-arm study of oral glutathione at 250-500 mg/day showed a significant reduction in the melanin index versus placebo.

    We tookMelanin-index reductions provide an objective pigmentation benchmark against which noova can calibrate its pigmentation/skin-tone-evenness scoring over time.

  22. Efficacy of topical vitamin C in melasma and photoaging: A systematic review

    Correia G, et al. · Journal of cosmetic dermatology, 2023 · view paper

    FoundA systematic review of seven publications (139 volunteers) found significant lightening of skin treated with topical vitamin C on objective pigmentation assessments.

    We tookObjective pigmentation-lightening evidence supports calibration of noova's overall pigmentation score for expected treatment-driven change.

Product match

14 papers

The measured effect sizes behind the actives we grade products on: retinoids, niacinamide, vitamin C, acids, ceramides, sunscreen.

  1. The Effect of Local Hyaluronic Acid Injection on Skin Aging: A Systematic Review and Meta-Analysis

    Zhou R, et al. · Journal of cosmetic dermatology, 2025 · view paper

    FoundRandom-effects meta-analysis found hyaluronic acid injection significantly improved skin hydration (SMD = 1.34) and radiance (SMD = 0.51) versus controls.

    We tookAnchors a strong (large-effect) product-match for HA on the hydration/radiance signals; weaker/insignificant for elasticity, so grade HA highest for hydration.

  2. Oral Hyaluronic Acid Supplement: Efficacy in Skin Hydration, Elasticity, and Wrinkle Depth Reduction

    Amin P, et al. · Journal of drugs in dermatology : JDD, 2025 · view paper

    FoundA meta-analysis of 7 RCTs of oral hyaluronic acid showed statistically significant improvements in skin hydration, elasticity, and wrinkle depth.

    We tookSupports HA product-match on hydration/wrinkle signals with pooled RCT evidence, though non-significant for firmness and TEWL (moderate confidence).

  3. Efficacy of topical treatments for mild-to-moderate acne: A systematic review and meta-analysis of randomized control trials

    Kakpovbia EE, et al. · Journal of the European Academy of Dermatology and Venereology : JEADV, 2025 · view paper

    FoundIn a network meta-analysis, adding clindamycin to tretinoin did not significantly improve non-inflammatory lesion reduction (RoM 1.13; 95% CI 0.94-1.36), so tretinoin drives the effect.

    We tookGrounds high product-match for adapalene/tretinoin combinations on acne/clarity; retinoid is the active component, so retinoid-containing acne products score strongly.

  4. Comparing Tretinoin to Other Topical Therapies in the Treatment of Skin Photoaging: A Systematic Review

    Siddiqui Z, et al. · American journal of clinical dermatology, 2024 · view paper

    FoundAcross 23 comparison studies, comparators to tretinoin were equivalent in 13, greater in 7, and less effective in 3, confirming tretinoin as the gold-standard anti-aging benchmark.

    We tookEstablishes tretinoin as the reference product-match for photoaging/wrinkles; retinol/retinaldehyde products grade near-equivalent with better tolerability.

  5. Topical azelaic acid, salicylic acid, nicotinamide, sulphur, zinc and fruit acid (alpha-hydroxy acid) for acne

    Liu H, et al. · The Cochrane database of systematic reviews, 2020 · view paper

    FoundIn a Cochrane review, azelaic acid was slightly less effective than benzoyl peroxide (RR 0.82) but comparable to tretinoin (RR 0.94) for acne treatment response.

    We tookGrounds azelaic acid product-match for acne/clarity at roughly tretinoin-level efficacy; the review also covers salicylic acid and glycolic peels for the AHA/BHA signals.

  6. Use of sunscreen and risk of melanoma and non-melanoma skin cancer: a systematic review and meta-analysis

    Silva ESD, et al. · European journal of dermatology : EJD, 2018 · view paper

    FoundA meta-analysis of 29 studies (313,717 participants) found no significant association between sunscreen use and skin cancer (OR 1.08; 95% CI 0.91-1.28).

    We tookCalibration caution for sunscreen product-match: observational data show no population-level protection signal, so avoid over-weighting SPF match on cancer-prevention grounds.

  7. Comparative efficacy of topical interventions for facial photoaging: a network meta-analysis

    Lin L, et al. · Scientific reports, 2025 · view paper

    FoundA network meta-analysis of 23 RCTs (3905 participants) ranked isotretinoin, retinol, and tretinoin as significantly improving fine wrinkles, with glycolic acid best for roughness.

    We tookProvides relative product-match ranking for photoaging: retinoids top for wrinkles/hyperpigmentation, glycolic acid for texture/roughness, informing signal-specific weighting.

  8. Effect of Nicotinamide in Skin Cancer and Actinic Keratoses Chemoprophylaxis, and Adverse Effects Related to Nicotinamide: A Systematic Review and Meta-Analysis

    Mainville L, et al. · Journal of cutaneous medicine and surgery, 2022 · view paper

    FoundNicotinamide (niacinamide) reduced skin cancer incidence versus control with a rate ratio of 0.50 (95% CI 0.29-0.85) across 5 trials.

    We tookSupports niacinamide %-match; a ~50% relative reduction in skin cancers is a strong efficacy signal for niacinamide's protective/actinic value.

  9. Interventions for photodamaged skin

    Samuel M, et al. · The Cochrane database of systematic reviews, 2005 · view paper

    FoundTopical 0.05% tretinoin improved photodamaged skin versus placebo with a relative risk of improvement of 1.73 (95% CI 1.39-2.14) at 24 weeks.

    We tookAnchors retinoid %-match for photoaging/wrinkles; RR 1.73 improvement confirms strong topical tretinoin efficacy (topical ascorbic acid also showed benefit).

  10. Topical supplementation with physiological lipids rebalances the stratum corneum ceramide profile and strengthens skin barrier function in adults predisposed to atopic dermatitis

    Andrew PV, et al. · The British journal of dermatology, 2025 · view paper

    FoundA topical multivesicular emulsion with physiological lipids significantly increased stratum corneum ceramide NP(18) by 24% and AP(18) by 19%.

    We tookDirect ceramide %-match evidence: a ceramide-lipid moisturizer measurably rebuilds SC ceramides (+24%/+19%) and improves barrier integrity.

  11. Efficacy of combined 1927 nm thulium fiber laser and supramolecular salicylic acid in photoaging treatment

    Qiao S, et al. · Lasers in medical science, 2025 · view paper

    FoundAdding 30% supramolecular salicylic acid to thulium laser reduced photoaging scores by 10.51% versus 7.26% for laser alone (P<0.001).

    We tookSupports salicylic acid (BHA) %-match for photoaging; adjunct SSA adds a measurable ~3% extra photoaging-score reduction.

  12. Barrier films or dressings for the prevention of acute radiation dermatitis in breast cancer: a systematic review and network meta-analysis

    Wong HCY, et al. · Breast cancer research and treatment, 2024 · view paper

    FoundBarrier films/dressings reduced moist desquamation versus standard care: Hydrofilm OR=0.08, Mepitel Film OR=0.31, StrataXRT OR=0.22.

    We tookBarrier/occlusive product evidence: strong odds-ratio reductions (OR 0.08-0.31) inform %-match for barrier-protective formulations.

  13. Skin care interventions in infants for preventing eczema and food allergy

    Kelleher MM, et al. · The Cochrane database of systematic reviews, 2022 · view paper

    FoundEmollient skin-care in infancy did not reduce eczema risk (RR 1.03, 95% CI 0.81-1.31) and probably increased skin infection risk.

    We tookCaps %-match for emollients as eczema PREVENTION (null RR 1.03); barrier moisturizers don't prevent eczema in healthy infants.

  14. The Impact of an AGE-Inhibiting Moisturizer on Procedure Effectiveness

    Draelos ZD, et al. · Journal of drugs in dermatology : JDD, 2025 · view paper

    FoundAfter a glycolic acid peel, an AGE-inhibiting moisturizer beat control on 5 facial attributes (clarity, tone, fine lines, elasticity, overall appearance) at week 8 (all P<0.05).

    We tookPost-peel adjunct moisturizer %-match; significant gains on 5 aging attributes (P<0.05) support pairing glycolic (AHA) procedures with a barrier moisturizer.

Calibration

12 papers

The honesty layer: how much a phone camera can recover versus clinical instruments, and why photo quality has to gate the score.

  1. Development and validation of neonatal jaundice detection using smartphone application

    Shen GT, et al. · Computers in biology and medicine, 2025 · view paper

    FoundA smartphone jaundice model using a color-calibration card identified 89.4% of newborns with TsB>15 mg/dL at an AUC of 0.963, sensitivity 90.1% and specificity 89.0%.

    We tookDemonstrates that smartphone capture with an explicit color-calibration card reaches near-clinical accuracy, supporting noova's image-quality gating and standardized-capture requirements.

  2. Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists

    Haenssle HA, et al. · Annals of oncology : official journal of the European Society for Medical Oncology, 2018 · view paper

    FoundA CNN's ROC AUC exceeded the mean AUC of 58 dermatologists (0.86 versus 0.79, P < 0.01) for dermoscopic melanoma classification.

    We tookBenchmarks AI-vs-dermatologist agreement, informing how much confidence noova can assign to model-derived skin assessments relative to expert judgment.

  3. Dermatologist-level classification of skin cancer with deep neural networks

    Esteva A, et al. · Nature, 2017 · view paper

    FoundA single CNN trained on 129,450 clinical images spanning 2,032 diseases matched 21 board-certified dermatologists on skin-cancer classification.

    We tookEstablishes the dataset scale needed for dermatologist-level image classification, a reference point for calibrating noova's model confidence and smartphone-delivered scoring.

  4. Artificial intelligence in dermatology: advancements and challenges in skin of color

    Fliorent R, et al. · International journal of dermatology, 2024 · view paper

    FoundOnly 30% of the AI dermatology programs reviewed reported any data on their use in skin of color, and current programs performed worse at identifying lesions in skin of color.

    We tookQuantifies the skin-of-color evidence gap driving fairness guardrails, cautioning that noova's scores may be less reliable on darker skin tones without SOC validation.

  5. Artificial intelligence for nonmelanoma skin cancer

    Trager MH, et al. · Clinics in dermatology, 2024 · view paper

    FoundOnly three of the reviewed publications addressed skin of color and AI for nonmelanoma skin cancer, flagging limited diverse training data.

    We tookReinforces the scarcity of skin-of-color evidence, supporting noova's fairness guardrails and skepticism toward model generalizability across skin tones.

  6. AI in Psoriatic Disease: Scoping Review

    Barlow R, et al. · JMIR dermatology, 2024 · view paper

    FoundA scoping review of 38 papers found AI applied to psoriatic disease, including teledermatology diagnosis from patient-taken photographs and severity estimation.

    We tookConfirms patient-captured photos can drive severity estimation, relevant to noova's image-based severity scoring while noting validation and skin-phototype generalizability limits.

  7. Diagnostic reliability in teledermatology: a systematic review and a meta-analysis

    Bourkas AN, et al. · BMJ open, 2023 · view paper

    FoundAcross 44 studies, teledermatology vs face-to-face pooled diagnostic agreement was 68.9% with kappa 0.67, rising to 71% (kappa 0.69) for dermatologists but only 44% (kappa 0.52) for non-specialists.

    We tookAnchors an expected image-based-vs-in-person agreement ceiling (~0.67 kappa) and shows rater expertise plus image-acquisition training swing accuracy sharply, informing noova's confidence gating.

  8. Comparative and complementary diagnostic value of dermatoscopy and clinical close-up photography in skin cancer diagnosis: A study from the MILK10k dataset

    Müller C, et al. · Journal of the American Academy of Dermatology, 2026 · view paper

    FoundIn a 283-reader study on 1567 paired images, dermatoscopy alone had 85.0% sensitivity vs 74.2% for clinical close-ups, but lower specificity (66.8% vs 71.9%).

    We tookQuantifies the sensitivity/specificity tradeoff of consumer-style close-up photos vs dermatoscopy, calibrating how much diagnostic weight noova can place on a phone image lacking dermatoscopic detail.

  9. Diagnostic Accuracy of Pediatric Teledermatology Using Parent-Submitted Photographs: A Randomized Clinical Trial

    O'Connor DM, et al. · JAMA dermatology, 2017 · view paper

    FoundConcordance between parent smartphone photos and in-person diagnosis rose to 89% (kappa 0.88) when photos were of high enough quality, vs 83% (kappa 0.81) overall.

    We tookDirect evidence that filtering to high-quality smartphone images materially lifts diagnostic concordance, supporting noova's image-quality gate before scoring.

  10. Diagnostic Reliability of In-Person Versus Remote Dermatology: A Meta-Analysis

    Bastola M, et al. · Telemedicine journal and e-health : the official journal of the American Telemedicine Association, 2021 · view paper

    FoundMeta-analysis of six dual-in-person studies found remote diagnoses significantly less concordant than in-person (OR = 0.55, 95% CI 0.42-0.72, I2 = 58%).

    We tookEstablishes that remote/image-only assessment is systematically less reliable than in-person, a prior for capping noova's confidence on photo-only scans of risky lesions.

  11. A longitudinal study of consistency in diagnostic accuracy of teledermatology tools

    Kanthraj GR, et al. · Indian journal of dermatology, venereology and leprology, 2013 · view paper

    FoundAggregated over 15 years, store-and-forward teledermatology averaged 73.35% +/- 14.87% diagnostic accuracy vs 70.37% +/- 7.01% for video, both stable against face-to-face gold standard.

    We tookGives a stable long-run accuracy baseline (~73%) for asynchronous image-based diagnosis, useful for calibrating noova's expected agreement and confidence bands.

  12. Teledermatology and in-person examinations: a comparison of patient and physician perceptions and diagnostic agreement

    Lowitt MH, et al. · Archives of dermatology, 1998 · view paper

    FoundAmong 139 patients, agreement between live video teledermatology and in-person diagnoses was 80%.

    We tookEarly benchmark that remote-vs-in-person diagnostic agreement plateaus around 80%, informing a realistic upper bound for noova's photo-based confidence.

Now see it read your skin.

One text when we open your spot. Nothing else, ever.