Diagnostic and Personalized Skin Care via Artificial Intelligence Ankur Purwar PhD Principal Scientist, Procter & Gamble R&D, Singapore
10 research centers 65 OLAY has been creating 1,000 scientifically-advanced researchers Skin Care for years Over 80 consumers every year Million
67% Only 67% of women were able to find what they were looking for when browsing Satisfied the facial Skin Care aisle CONFUSION 14% 14% of women say they don’t know what their specific Skin Don ’ t know Care needs are 1 1 P&G data on file
OLAY SKIN ADVISOR Your Personalized Skin Care regimen in a snap
DISCOVERY PROOF DIAGNOSTICS Discoveries leading Personalized Skin Care Empowering skincare now to OLAY Skin Advisor in a snap and in the future
DISCOVERY
Discoveries Leading to OLAY Skin Advisor Imaging Facial Mapping Breakthroughs Study VizID ™ Learning Algorithm Platforms MDE and MES Studies
Pioneers in skin imaging for 25 years+ Chromophore OLÉ ™ OLAY Skin Imaging for Beauty Imaging VISIA ™ mapping and (Overhead Lighting Advisor R&D System technology SIAScopy Environment)
Multi-Decade Ethnicity (MDE) Genomic Study Actual Age 44 years Caucasian African Perceived: Perceived Age 57 years Prof Alexa Kimball Perceived: 29 years Identification of “Exceptional Agers”. Real (Chronological) Age
Multi-Ethnic Skin (MES) Phenotypic Study Canfield VECTRA Full-face 3D capture Characterisation of the facial skin of the world’s women.
Ageing Appearance Prediction Models Topography Shape Colour Original Image corrected to corrected to corrected to average average average
Ageing appearance simulation – what’s possible!
Facial Mapping Study Forehead Understanding the differential characteristics and ageing of skin Crow’s in five zones across the face. Feet Under Eye Flagler, M et al., New biological insights into skin ageing around the eye. Presented at the 74 th Cheek Annual Meeting of the American Academy of Dermatology, Washington D.C., 2016 Chin
VizID ™ Learning Algorithm Platforms
Visible Skin Age Prediction Feature Detection Age Prediction Predicted Age Trained a deep learning model with 50,000 facial images ▪ Model is able to further detect a person’s key aging areas and determine how old they ▪ actually look.
Technical Validation of Skin Age Model More precise than dermatologist’s at predicting visible age Derm Range= Visage Range= 20.27 years 2.88 years Image Age Range Visage Dermatologist N=625 people N=4977 total derm (model sees same person 8 times) (8 different derms see same person 1 time) grades Weitz, S et al., Improving Consumer Compliance Through Better Product Recommendation- New Skin Advisor Tool. Presented at the 75 th Annual Meeting of the American Academy of Dermatology, 2017
DIAGNOSTICS
It Starts with a Scientific ‘Selfie’ Step 2 Step 1 Step 4 Step 3 Step 4 Product Skin Analysis Recommendation Image Acquisition Questionnaire Regimen Reco
Computed “Skin Age”. “Best” and “Improvement Needed” areas identified.
Regimen recommendation augmented by a “Synaptic Intelligence” platform from Nara Logics, Inc.
PROOF
Results 1. 100 US women, age 25-65, facial moisturizer users, were enrolled in a 4-week online consumer test. 2. Group 1 (n=50) received a product regimen based on the skin advisor deep learning algorithm and preferences and Group 2 (n=50) self-selected a product regimen. 3. Self-assessment questions were completed pre-use and post- 4 weeks product use. Key results metrics pre-product use indicate Key results metrics post- 4 weeks product use indicate satisfaction with the skin advisor product satisfaction with the skin advisor product recommendation and improved consumer recommendation compliance. Weitz, S et al., Improving Consumer Compliance Through Better Product Recommendation- New Skin Advisor Tool. Presented at the 75 th Annual Meeting of the American Academy of Dermatology, 2017
The Future: Starts Smart, Gets Smarter 2018 and Beyond System upgrades Pioneering new “smart” technology Smarter with every use…
OLAY Skin Advisor Your Personalized Skin Care regimen in a snap
Recommend
More recommend
