Prologue — In 2026, the Mirror Makes a Diagnosis
In 2019, a beauty store was a place where a friendly associate brushed foundation onto the back of your hand and said "Shade 21 suits you." In 2026 it is different. A user steps in front of a mirror, a camera scans the face, color matching happens in LAB color space, pores and wrinkles and sebum and redness are analyzed pixel-by-pixel, and AR overlays makeup on top.
Dermatology clinics have also changed. In 2024 the FDA granted the first point-of-care skin cancer detection clearance. A primary care physician can scan a patient's mole in 30 seconds and assess melanoma risk. The same year, the EU AI Act classified facial data as "high-risk biometrics."
The 2026 landscape.
- SkinVision · DermaSensor · VisualDx · Aysa · MoleScope · Miiskin — medical-grade skin AI has moved from the patient's hand into the clinic.
- ModiFace (L'Oréal acquisition) · Sephora Virtual Artist · YouCam Makeup (Perfect Corp) — AR makeup is now a mobile baseline.
- Korean Glowpick · Hwahae · Maskpect · Loon Lab · Doctor Inside · CHEAFY · Olive Young · Musinsa Beauty — the digital infrastructure of K-beauty.
- Japanese @cosme · LIPS · SHISEIDO Optune · POLA APEX · Shiseido Benefique · Orbis · Kao RECORE · ESTEE LAUDER iMatch — the quiet AI penetration of J-beauty.
- Lumini Plus · HiMirror · Canfield VISIA — skin scanners are expanding both into homes and into clinics.
The skin is no longer "how your face looks today." It is measured, tracked, recommended on. This piece maps that whole terrain.
One-line summary: "Whose face, on what camera, with what labels, sent where, judged by whom." Those five questions decide 90% of beauty AI choices.
Chapter 1 · Why Beauty AI Exploded in 2026
The beauty AI value proposition is simple.
- Conversion rate — Customers who can see their shade buy. Average lift +30–50%.
- Return rate — Color cosmetics were a 25%-return category in Korea; AR matching brings it down by up to -15pp.
- Time on site — VTM (Virtual Try-Makeup) pages average 4–6x the dwell time of normal PDPs.
- Medical side — Primary care melanoma referrals dropped by 30% with triage tools.
Three technical tailwinds converged.
- Mobile cameras — not dermatoscope-grade, but good enough.
- Diffusion / GAN — realistic makeup synthesis became practical.
- CNNs + ViTs — skin lesion classification approached board-certified dermatologist accuracy.
Between 2024 and 2026, beauty and skin AI crossed from "demo" to "shipped feature."
Chapter 2 · SkinVision — The Veteran of Melanoma Triage
SkinVision is a melanoma-triage app headquartered in Amsterdam. A user photographs a mole or skin lesion with a smartphone and within 30 seconds receives a risk rating (high / medium / low) and a next-step recommendation.
Key facts.
- CE Mark Class IIa medical device certification.
- Free access in about nine European countries and Australia via insurer / health-plan partnerships.
- Around 2M+ cumulative users.
- Reported melanoma detection sensitivity around 95%.
- Used by some UK NHS GPs as a pre-referral triage tool.
SkinVision's business model is B2B2C, not direct consumer — it is embedded into insurer, pharma, and health-system offerings. There is no US FDA clearance; the operation is Europe-centric.
The accurate framing of this category is not "AI diagnoses" but "AI triages physician referrals." Final diagnostic responsibility still sits with a dermatologist.
Chapter 3 · DermaSensor — FDA-Cleared Point-of-Care Skin Cancer Detection
DermaSensor received an FDA De Novo clearance in January 2024 for non-invasive skin cancer detection. Where SkinVision lives in the patient's hand, DermaSensor lives in the primary care physician's hand.
How it works.
- A prism optical module is held against a suspicious lesion and elastic-scattering spectroscopy is run for about 30 seconds.
- AI analyzes the spectral pattern and classifies the lesion as "Investigate Further (refer)" or "Monitor."
- All three major skin cancers — melanoma, basal cell carcinoma, squamous cell carcinoma — are in scope.
The DERM-SUCCESS clinical study reported around 96% sensitivity for melanoma and around 96% across all skin cancers. The American Academy of Family Physicians (AAFP) endorsed its use in primary care.
DermaSensor's real value shows up in dermatologist-shortage areas. In the US, average dermatologist wait times exceed 30 days in many counties. Fast triage in primary care means the patients who actually need a specialist get there faster.
Chapter 4 · VisualDx · Aysa · MoleScope · Miiskin — Adjacent Medical Skin AI
Beyond SkinVision and DermaSensor, the medical skin AI ecosystem is layered.
- VisualDx — A clinical decision support (CDS) tool running since 1999. Clinicians enter a patient photo and symptoms and receive a differential diagnosis. Database of 2,500+ conditions. Adopted by about 80% of US medical schools.
- Aysa (Visual.AI) — A patient-facing skin checker. Image plus free text returns 5–10 candidate categories and prompts a dermatologist visit. Not a diagnostic device — an information tool that tells you what kind of condition this might be.
- MoleScope — A smartphone-attachable dermatoscope module for high-resolution mole imaging. A dermatologist's follow-up tool.
- Miiskin — A mole-tracking app. The user re-photographs the same spot over time and the app visualizes the change. The ABCDE rule (Asymmetry, Border, Color, Diameter, Evolving) is built into the educational content.
The shared message — AI does not replace dermatologists. It shortens the cycle.
Chapter 5 · ModiFace — The AR Makeup Backbone After the L'Oréal Acquisition
ModiFace was a 2007 University of Toronto spinout in AR beauty, acquired by L'Oréal in 2018. The acquisition price was not disclosed, but it was the beauty industry's first AR acquisition and drew significant attention.
Core ModiFace technology.
- Real-time face tracking and landmark detection.
- Lipstick · eyeshadow · eyeliner · blush · foundation · hair color simulation.
- Skin tone matching and foundation shade recommendation.
- Daily makeup routine guidance.
Post-acquisition, ModiFace tech is embedded in the proprietary apps and e-commerce of L'Oréal Paris · Lancôme · YSL Beauty · Maybelline · NYX · Garnier. ModiFace-powered try-on also runs inside Amazon, Facebook (Meta), KakaoTalk, and LINE.
A standard pattern emerged — brands stopped building AR from scratch and started calling the ModiFace API instead, avoiding the cost of an in-house computer vision team while keeping quality.
Chapter 6 · Sephora Virtual Artist — The US Retail Standard
Sephora Virtual Artist launched in 2016 as Sephora's AR makeup feature, running both inside the mobile app and on in-store kiosks. The technology started external but is now a mixed in-house + partner stack.
Key specs.
- iOS / Android app integration.
- 30K+ SKUs tryable, including lipsticks, eyeshadows, and false lashes.
- "Color Match" — extracts foundation tone from a selfie in LAB color space.
- "Skincare IQ" — personalized recommendations from skin type, tone, and concerns.
- In-store kiosks (Beauty Hub / Color IQ) — live try-on assisted by associates.
Sephora's try-on data trickles into the personalization models of other LVMH-portfolio brands (Dior, Givenchy, Fenty Beauty). It is effectively the case that defined the US beauty e-commerce AR baseline.
Chapter 7 · YouCam Makeup · Perfect Corp — The Asian-Origin AR Leader
Perfect Corp is a Taiwan-headquartered beauty AR specialist that runs YouCam Makeup · YouCam Perfect · YouCam Nails · YouCam Hair · YouCam Salon apps and B2B SDKs. Cumulative app downloads exceed 1 billion.
Core assets.
- YMK (YouCam Makeup) — Consumer-facing free makeup try-on app. Number-one or near-number-one in Japan, Korea, Southeast Asia, and China.
- AgileFace — In-house face landmark SDK.
- B2B solutions — Partnerships with 200+ brands including Estée Lauder, Clinique, MAC, Shu Uemura, L'Oréal Paris, and NYX.
- AI Skin Diagnostic — 14-item selfie analysis covering pores, wrinkles, spots, dark circles, and more.
- Hair Color Try-On — Hair color synthesis.
- 3D Nail Try-On — Hand-tracking nail simulation.
The company went public on NYSE via SPAC merger in 2020 (ticker PERF). Over 2024–2026 it has added generative AI integrations (GenAI Beauty Advisor), giving it a chatbot-style makeup recommendation. In Korea it is the most-frequently-licensed backbone alongside ModiFace.
Chapter 8 · Glowpick · Hwahae — K-Beauty's Digital Infrastructure
Two companies effectively split Korean beauty e-commerce's digital infrastructure.
Glowpick — A cosmetics review platform launched in 2014. Around 10M+ cumulative members and 10M+ reviews. Category rankings, skin-type-based recommendations, and ingredient guides are all in scope. Recent AI recommendation upgrades learn the user's skin type, concerns, and preferences to personalize curation.
Hwahae (translates roughly as "deciphering cosmetics") — Launched in 2013. Holds roughly 100K+ products and 120K+ ingredients in its database. The signature feature is EWG ingredient grading — each ingredient's hazard rating is auto-displayed. Combined with a global cosmetics database at the 1B+ SKU scale, the "users of this product also like" recommendation also works well.
Difference in flavor.
- Glowpick — Review and ranking centric. "What others love."
- Hwahae — Ingredient and safety centric. "What suits my skin."
K-beauty users typically have both apps installed and cross-reference. The general user view is that Hwahae's recommendation accuracy is higher — ingredient-based embeddings are less noisy than review-text-based ones.
Chapter 9 · Maskpect · Loon Lab · Doctor Inside · CHEAFY — K-Beauty AI's Next Generation
If Glowpick and Hwahae are gen 1, then 2024–2026 saw the rise of gen 2 K-beauty AI.
- Maskpect — A sheet mask recommendation app, specialized because Korea is the home of the sheet mask. The user's daily condition and skin concerns lead to a matched mask recommendation.
- Loon Lab — Femtech plus skin AI. Learns the correlation between menstrual cycle and skin state and recommends care per cycle phase. Hormonal swings, acne, dryness, and sensitivity are mapped together.
- Doctor Inside — A telemedicine dermatology consultation platform. The user uploads a photo, the AI does first-pass classification and matches to a suitable dermatologist, with consult payment and prescription issuance in the same flow.
- CHEAFY — A Korean-style dermatology consult plus prescription app, similar to Doctor Inside but with stronger focus on chronic tracks like acne and pigmentation.
- GoodDoc Derma — GoodDoc's dermatology category. Telemedicine plus AI triage.
Korea's strength in this category is the medical law landscape. The Korean telemedicine pilot expanded in 2024 and dermatology is classified as a high-fit category for remote consults. AI triage plus physician judgment plus prescription delivery flows together in one pipeline.
Chapter 10 · Olive Young · Musinsa Beauty · Sephora Korea — K-Beauty Retail AI
The retail side has evolved fast.
- Olive Young — From 2024, integrated proprietary AI skin analysis into store kiosks and the app. Leverages CJ ENM data infrastructure. A single in-store selfie produces a skin score and a recommended product list.
- Musinsa Beauty — The beauty category of Musinsa. Visual search (find similar products from an image), personalized recommendation, and AR try-on (via partners).
- Sephora Korea — Tuned the global Virtual Artist for the Korean market, with additional Korean skin tone data.
- Lotte ON Beauty — Lotte Group's AI recommendation.
- CJ Olive Young + AI shelving — In-store displays change dynamically by user.
The shared K-beauty retail pattern is single selfie → skin score → recommended products. The selfie has hardened as the entry point, the measurement varies by brand, and the user experience has standardized.
Chapter 11 · AmorePacific · LG Household & Health Care · AIDA Lab — Korean Conglomerate AI
The K-beauty conglomerates also run in-house AI teams.
- AmorePacific AI Magic Mirror — In-store digital mirror. Step in front and skin analysis, tone matching, and makeup simulation auto-trigger.
- AmorePacific ARI (AmoreCare Real Image) — Proprietary skin analysis model. UV damage, spots, pores, wrinkles, and sebum across multiple dimensions.
- LG Household & Health Care + AIDA Lab — LGH&H's AI Beauty research arm. Owns the recommendation algorithms for Whoo, Su:m37, O HUI, and Belif lines.
- LG CNS Beauty AI — Group IT subsidiary operating beauty-category AI solutions.
- Hyundai Department Store The Handsome Beauty — Luxury department store AI curation.
Conglomerate AI lives not in a single app but across all lines' recommendation, CRM, and CDP. The premise is that one user uses multiple brands from the portfolio, so cross-line recommendation is the core.
Chapter 12 · @cosme · LIPS — Japan's Two Pillars of Beauty Reviews
Japan's beauty digital infrastructure has two pillars.
@cosme (アットコスメ) — Started in 1999, it is Japan's largest cosmetics review site. Roughly 18M+ cumulative reviews and 300K+ products in the database. Offline stores @cosme TOKYO and @cosme STORE are also run.
Core elements.
- User reviews plus expert ratings plus category rankings.
- Recommendations by skin type and age band.
- "@cosme BEST COSMETICS AWARDS" — the de facto Japanese beauty awards each year.
LIPS — A Gen Z-focused beauty review and video app launched in 2018. Where @cosme is text-review-centric, LIPS is short-form makeup tutorial and review video.
- 15M+ cumulative downloads.
- Trend discovery via influencers.
- AI recommendation combines text and video embeddings.
Unlike Korea, Japan still has strong in-store BA (Beauty Advisor) counseling culture, so digital and offline blend more visibly. @cosme STORE BAs reference the user's @cosme app data when making in-store recommendations.
Chapter 13 · SHISEIDO Optune · POLA APEX · Shiseido Benefique — Japan's Personalization Camp
J-beauty is the global leader in the "luxury personalization" track.
- SHISEIDO Optune — Launched in 2017. Every morning the user inputs skin state via selfie plus weather and mood; a dedicated dispenser brews the day's skincare from 100K+ combinations. Subscription model. Attempted global expansion in 2024.
- POLA APEX — Five generations of evolution since 1989. Roughly 18M+ cumulative consultation records power this ultra-personalized luxury skincare. In-store IoT devices measure skin and a formulation is chosen from approximately 5,000+ combinations.
- Shiseido Benefique 肌診断 — Benefique's in-store kiosk skin diagnosis. Pores, wrinkles, UV damage in multi-dimensional view.
- Orbis + AI 肌診断 — Orbis's mobile-app skin diagnosis. Single selfie analysis.
- Kao RECORE + AI brightening — Kao's brightening-line AI matching.
- POLA RED B.A — Digital counseling for the luxury line.
POLA APEX's real edge is decades of Japanese skin data. Domain knowledge of East Asian skin (especially UV damage and pigmentation patterns) is more precise than the generic global model of foreign luxury brands.
Chapter 14 · ESTEE LAUDER iMatch · L'Oréal MyShade — Global Luxury AI
Global luxury beauty AI integrations.
- ESTEE LAUDER iMatch Virtual Foundation Finder — A foundation matching tool that runs on Estée Lauder, MAC, Bobbi Brown, and Clinique proprietary sites, with some of the underlying tech provided by ModiFace.
- L'Oréal MyShade — Selfie-based foundation recommendation for Maybelline and L'Oréal Paris.
- Dior Backstage Mirror — Dior in-store kiosk.
- Fenty Beauty Pro Filt'r Match Stix — 50+ shade matching, strong on diverse skin tones.
- Charlotte Tilbury Magic Mirror — In-store digital mirror.
- Lancôme Le Teint Particulier — On-the-spot foundation formulation in-store. 70,000 combinations.
This category is less "AI does the matching" and more "AI removes the social friction of matching." It solves the problem of an associate not feeling able to say "that shade doesn't suit you" by letting the camera be objective.
Chapter 15 · Skin Scanners — Lumini Plus · HiMirror · Canfield VISIA
Camps of consumer and clinical skin scanners.
- Lumini Plus (Lululab) — A Korean Lululab clinical skin scanner. Multi-modal imaging in UV, PL (parallel polarized), CP (cross polarized), and WL (white light). Around 10 auto-analyzed features. High penetration in Korean dermatology and aesthetic clinics.
- HiMirror — Perfect Corp's smart mirror. Home use. Tracks skin change via morning selfies.
- Canfield VISIA — A medical-grade skin analysis system running since 1998. The de facto US dermatology standard. RBX technology separates hemoglobin and melanin channels.
- Aurora Glow Touch Up — Korean in-store digital counseling device.
- WAYB Pet Skin — Korean pet-skincare market's dog-skin scanner.
Consumer scanners run off the selfie camera and have limited precision; clinical scanners use a standardized light box, so longitudinal comparison is feasible. That is the decisive split.
Chapter 16 · Virtual Makeup Try-On — PicCollage · CyberLink BeautyCircle and Friends
Beyond ModiFace and Perfect Corp, the makeup category has a fan of other solutions.
- PicCollage Beauty — Collage-app-origin beauty filters.
- CyberLink BeautyCircle — Perfect Corp's sibling brand (now separated). PC-based makeup simulation plus video editing.
- FaceApp — Russian-origin. Strong on aging, makeup addition, and hair style change. Controversial on user data collection.
- Lensa AI — Prisma Labs. Selfie-based avatars plus makeup generation.
- YouCam Perfect — Perfect Corp's selfie retouch plus makeup combo app.
- Snow / Soda / B612 — Snapchat-like camera apps with built-in makeup filters.
About 90% of consumers' perceived "AI selfie beautification" happens inside this category. "Make me look prettier" beats "diagnose me" on value proposition.
Chapter 17 · AI Hair and Scalp Diagnostics — Tricho.AI · Hello Tricho · Cellbio
Beyond skin, hair and scalp went digital quickly.
- Tricho.AI — AI analysis of trichoscopy video. Auto-measures density, thickness, and follicle cleanliness.
- Folliscope — First-generation trichoscopy video analysis.
- Hello Tricho — Korean-origin scalp diagnosis solution. Wide retail and salon adoption.
- Cellbio — Korean scalp-care device plus AI analysis.
- NUNYI — Korean hair-loss care app.
- Dr. Mo — Telemedicine hair-loss platform with prescriptions.
- Olly Hair / Hims for hair — US men's hair-loss care.
K-beauty's scalp market splits into "male hair loss" + "female scalp care" + "postpartum hair loss" tracks, and AI diagnosis is becoming the entry point for each.
Chapter 18 · Personalized Supplements — Care/of · Nourish3D · Hum Nutrition
The "Inner Beauty" adjacency also brought AI in.
- Care/of — A representative US personalized vitamin company. Acquired by Bayer in 2020 but shut down in September 2024. The market did not scale as expected.
- Persona Nutrition — Under Nestlé Health Science. Sunset in 2024.
- Nourish3d — Daily vitamins extruded as a 3D-printed gummy. UK / US.
- Hum Nutrition — Selfie-based skin diagnosis plus supplement recommendation. Actively operating.
- Persona — Nestlé's daily vitamins.
- Olly · Ritual — Mainstream-market vitamins with integrated lifestyle counseling.
- Korean NutriCore · FitLab — Korean personalized supplements.
The 2024 sunsets of Care/of and Persona Nutrition signaled "personalized vitamins as a standalone business is hard." The flow afterward has been integration into larger beauty / nutrition corporate lines.
Chapter 19 · AI Telederm Consultation — MDhair · First Derm · Doctor Inside · Dermatalk
Dermatology is one of the highest-fit telemedicine categories.
- MDhair — US hair telemedicine plus AI triage.
- First Derm — US dermatology telemedicine. Photo plus text in, board-certified dermatologist response in 48 hours.
- DermatologistOnCall — First-generation US derm telemedicine.
- Doctor Inside · GoodDoc Derma · CHEAFY — Korean derm telemedicine.
- Dermatalk (デルマトーク) — Japanese derm telemedicine.
- Kry / Doctolib — European telemedicine, dermatology category strong.
The business models are typically two — per-consult fee (US$30–75) or monthly subscription (US$20–40/mo). The role of AI is primary triage plus physician time-savings. The diagnosis itself is done by a doctor.
Chapter 20 · In-Store AI Digital Mirrors — Magic Mirror · Beauty Hub · Smart Mirror
In-store digital mirrors became standard equipment in 2026 K-beauty and J-beauty.
- AmorePacific AI Magic Mirror — User recognition, skin analysis, makeup simulation, recommendation, and QR-pay in one flow.
- Sephora Beauty Hub — In-store kiosk version of Virtual Artist.
- HiMirror Smart Mirror — Perfect Corp consumer plus some retail.
- Coty Smart Mirror — Coty-portfolio brand stores.
- L'Oréal Connected Mirrors — Based on ModiFace tech.
- Shiseido Glow Mirror — Shiseido in-store digital counseling.
The real value of in-store digital mirrors is customer data collection. Visitor skin data and try-on patterns accumulate into the D2C CRM and feed back into online recommendation. Combined with the global stat that around 70% of AR try-on volume happens on mobile (not in store), the mirror becomes the "in-store experience → continue on mobile" entry point.
Chapter 21 · GLP-1 + Skin · AI Skin Age — 2026's Hot Trends
Three hot beauty AI trends in 2026.
GLP-1 + skin narrative — The prescription surge for Ozempic, Wegovy, and Mounjaro (GLP-1 agonists) is affecting the skin category. "Ozempic face" (volume loss from weight loss) became a digital-media talking point, growing the anti-aging and volume-restoration market. Some Korean beauty brands launched "post-GLP-1 skincare" lines.
AI skin-age estimation — Predicting skin age from a selfie became a standard feature. Confidence is around ±5 years. Strong as a marketing tool (users sharing "my skin age is 27" on social).
K-pop-based personalization — Korean idol skincare routines are being released as data. Member-by-member care libraries for NewJeans, LE SSERAFIM, and ATEEZ exist as separate categories inside Glowpick and Hwahae. "Same tone as my favorite idol" recommendation is the marketing engine.
All three trends share the same loop — AI measures → social amplifies → purchase. The same loop, repeated.
Chapter 22 · Privacy · Medical Law · EU AI Act — The Legal Weight of Facial Data
Facial data is not regular personal information. It is biometric. The legal weight is different.
- EU AI Act — In force since 2024. Classifies facial recognition / classification systems as "high-risk," requiring data governance, transparency, and human oversight. General application from 2026.
- GDPR Article 9 — Biometric data is a "special category," requiring explicit consent.
- Korea Personal Information Protection Act and Medical Service Act — Facial recognition is sensitive information. Whether an AI skin diagnosis counts as a "medical procedure" is contested. That ambiguity is why Doctor Inside and CHEAFY chose explicit physician-mediated models.
- Japan 個人情報保護法 and 医療広告 ガイドライン — Japan has strict medical-advertising rules. Non-medical lines like SHISEIDO Optune can advertise; calling AI skin analysis "medical" can violate the rule.
- US BIPA (Illinois) — Illinois Biometric Information Privacy Act. The legal basis for the roughly US$650M Facebook Tag Suggest settlement.
- FDA SaMD (Software as a Medical Device) — Any diagnostic / triage functionality is subject to FDA medical device clearance.
The deeper beauty / skin AI moves into the "medical" zone, the higher the regulatory cost. The 2026 average global beauty AI position is the safer zone of "wellness, recommendation, non-medical."
Chapter 23 · Beyond 2026 — The Next Five Years of Beauty AI
Direction of travel for the next five years.
- On-device inference — Selfie analysis on the mobile NPU. No cloud transit, stronger privacy.
- Diffusion makeup synthesis — GAN-era ModiFace will migrate toward diffusion. Another step up in realism.
- 3D face model plus AR — Face tracking on Apple Vision Pro and Meta Quest may displace home mirrors.
- DTx (Digital Therapeutics) skin — Prescription-form digital therapeutic apps enter the acne, atopic dermatitis, and psoriasis tracks.
- GenAI beauty advisor — Chatbot-style makeup coaches. Perfect Corp started in 2024 and L'Oréal Beauty Genius followed.
- Data partnerships — Sephora, Olive Young, and @cosme user data may surface as category statistics for external use.
- K-beauty global integration — Global expansion (or acquisition) of Glowpick and Hwahae becomes visible.
- Pet beauty — Pet skin AI like WAYB Pet remains small as a category but grows fast.
The common engine behind all this is the behavioral change that users measure their face daily. Once they start measuring, they do not stop. The market size of beauty AI is being built on top of that behavior.
Epilogue — Where to Start
The entry point depends on the use case.
- Consumer · daily selfie — Start with Glowpick, Hwahae, your whitening line, or YouCam Makeup. For K-beauty users, Hwahae is essential.
- Consumer · makeup try-on — Sephora Virtual Artist, YouCam Makeup, or brand apps with ModiFace inside.
- Consumer · skin tracking — Miiskin, HiMirror, AmorePacific Magic Mirror.
- Consumer · suspicious lesion — SkinVision (Europe), Aysa (US) → dermatologist referral.
- Clinic · primary care — DermaSensor, VisualDx.
- Clinic · dermatology — Canfield VISIA, Lumini Plus, in-house camera box.
- Brand · AR integration — Perfect Corp, ModiFace, Vue.ai, YouCam B2B.
- Brand · D2C personalization — Mass-custom in the SHISEIDO Optune mold.
- Developer · model training — HAM10000, ISIC Archive, DermNet NZ datasets.
The first question is always the same — "Is this medical, recommendation, or marketing?" The answer reshapes regulation, liability, data, and cost all at once. May this piece be useful at that fork.
Beauty AI lives in the gray zone between "look prettier" and "be healthier." Users who define that gray zone clearly end up choosing the best tools.
References
- SkinVision. "Skin cancer detection app." https://www.skinvision.com/
- DermaSensor. "FDA cleared point-of-care skin cancer detection." https://www.dermasensor.com/
- VisualDx. "Clinical decision support." https://www.visualdx.com/
- Aysa (Visual.AI). "Skin condition checker." https://www.aysa.ai/
- Miiskin. "Mole tracking app." https://miiskin.com/
- ModiFace (L'Oréal). "AR beauty tech." https://modiface.com/
- Sephora. "Virtual Artist." https://www.sephora.com/beauty/virtual-artist
- Perfect Corp. "YouCam apps and B2B AR." https://www.perfectcorp.com/
- Glowpick. "Glowpick cosmetics review." https://www.glowpick.com/
- Hwahae. "Hwahae — deciphering cosmetics." https://www.hwahae.co.kr/
- SHISEIDO Optune. "Personalized skincare." https://optune.jp/
- POLA. "APEX personalized." https://www.pola.co.jp/special/o/apex/
- @cosme. "Cosmetic review site." https://www.cosme.net/
- LIPS. "Beauty review app." https://lipscosme.com/
- Estée Lauder. "iMatch Virtual Foundation Finder." https://www.esteelauder.com/
- Lululab. "Lumini Plus skin scanner." https://lulu-lab.com/
- Canfield Scientific. "VISIA Skin Analysis." https://www.canfieldsci.com/imaging-systems/visia-skin-analysis-system/
- Doctor Inside. "Korean dermatology telemedicine." https://www.doctorinside.co.kr/
- EU AI Act. "Regulation (EU) 2024/1689." https://eur-lex.europa.eu/eli/reg/2024/1689/oj
- FDA. "DermaSensor De Novo Clearance." https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfPMN/denovo.cfm
- HAM10000 Dataset. "Skin lesion classification benchmark." https://www.nature.com/articles/sdata2018161
- ISIC Archive. "International Skin Imaging Collaboration." https://www.isic-archive.com/
- Olive Young Tech. "Olive Young AI skin analysis." https://www.oliveyoung.co.kr/
- Esteban et al. "AI for melanoma detection — Nature Medicine review." https://www.nature.com/articles/s41591-020-0942-0