From consistent subjective assessment of skin sensitivity severity to its accurate objective scoring.

BACKGROUND: Sensitive skin (SenS) is a syndrome leading to unpleasant sensations with little visible signs. Grading its severity generally relies on questionnaires or subjective ratings. MATERIALS AND METHODS: The SenS status of 183 subjects was determined by trained assessors. Answers from a four-item questionnaire were converted into numerical scores, leading to a 0-15 SenS index that was asked twice or thrice. Parameters from hyperspectral images were used as input for a multi-layer perceptron (MLP) neural network to predict the four-item questionnaire score of subjects. The resulting model was used to evaluate the soothing effect of a cosmetic cream applied to one hemiface, comparing it to that of a placebo applied to the other hemiface. RESULTS: The four-item questionnaire score accurately predicts SenS assessors' classification (92.7%) while providing insight into SenS severity. Most subjects providing repeatable replies are non-SenS, but accepting some variability in answers enables identifying subjects with consistent replies encompassing a majority of SenS subjects. The MLP neural network model predicts the SenS score of subjects with consistent replies from full-face hyperspectral images (R2 Validation set  = 0.969). A similar quality is obtained with hemiface images. Comparing the effect of applying a soothing cosmetic to that of a placebo revealed that subjects with the highest instrumental index (> 5) show significant SenS improvement. CONCLUSION: A four-item questionnaire enables calculating a SenS index grading its severity. Objective evaluation using hyperspectral images with an MLP neural network accurately predicts SenS severity and its favourable evolution upon the application of a soothing cream.

SCAWA scales: A new digital tool for wrinkles clinical grading based on AI.

OBJECTIVE: Clinical assessment of wrinkle depth is essential for efficacy evaluations of anti-ageing products. Standardized photographic scales, representative of different wrinkle depths are often used by experts to assign subjects reliable grades. These tools, based on real pictures, usually exist as hard copies (printed books or sheets) for in vivo gradings. Our project aims at developing a methodology to create digital standardized computer-generated scales, allowing photograph and real-life gradings, and providing raters with greater comfort, accessibility, and flexibility in their construction, thanks to the artificial intelligence significative contribution. METHODS: A completely new approach, based on machine learning, allows the creation of Standardized ColorFace® AI-based Wrinkle Assessment (SCAWA) scales. Instead of using real photographs, the scale images are computer-generated. A generative adversarial network (GAN) is trained to create realistic wrinkle samples that are finely controllable by exploring the GAN latent space. Finally, the scale images are selected among hundreds of artificial images depicting natural wrinkle appearances, such as the illustrated wrinkle evolution is well-detailed (small gaps between grades), morphologically stable, and mathematically linear according to a criterion of wrinkle conspicuous depth. RESULTS: The created 12-point scale for crow's feet wrinkle evaluation on ColorFace® pictures is proven to be realistic, linear, and robustly and accurately usable for photograph assessments. The scale coherence in terms of image ranking has been validated, as well as its reliability and acceptability in real conditions of use. Additionally, the wrinkle grades obtained by the SCAWA scale are well correlated (R = 0.94) with the ones obtained by the Skin Aging Atlas on the same pictures. The AI methodology and digital format brought also interesting side results, such as an enhanced harmonization between experts and a higher representativeness, that is, a decrease of out-of-range pictures. CONCLUSION: SCAWA scale makes the most of machine learning to provide an innovative digital tool to ease wrinkles in visual assessment of pictures, while optimizing linearity, homogeneity, and accuracy aspects. The experts' enthusiastic feedback about the scale format and quality is promising regarding the adaptation of the methodology to other signs and a larger distribution of this tool in the market of cosmetic product efficacy assessment.

OBJECTIF: L'évaluation clinique de la profondeur des rides est essentielle pour les évaluations de l'efficacité des produits anti­âge. Des échelles photographiques standardisées, représentatives de différentes profondeurs de rides, sont souvent utilisées par les experts pour attribuer des notes fiables aux sujets. Ces outils, basés sur des images réelles, existent généralement sous forme de copies papier (livres ou feuilles imprimées) pour les notations in vivo. Notre projet vise à développer une méthodologie pour créer des échelles numériques normalisées générées par ordinateur, permettant des évaluations sur photographies et des évaluations en contexte de vie réelle, ce qui offrirait aux évaluateurs un plus grand confort, une meilleure accessibilité et davantage de flexibilité dans leur construction, grâce à la contribution significative de l'intelligence artificielle. MÉTHODES: Une approche totalement nouvelle, basée sur l'apprentissage automatique, permet de créer des échelles standardisées d'évaluation des rides fondées sur l'IA ColorFace® (Standardised ColorFace® AI­based Wrinkle Assessment, SCAWA). Au lieu d'utiliser de vraies photographies, les images de l'échelle sont générées par ordinateur. Un réseau génératif antagoniste (generative adversarial network, GAN) est formé pour créer des échantillons de rides réalistes qui sont finement contrôlables en étudiant l'espace latent GAN. Enfin, les images de l'échelle sont sélectionnées parmi des centaines d'images artificielles représentant des aspects de rides naturelles, où l'illustration de l'évolution des rides est bien détaillée (petits écarts entre les grades), morphologiquement stable et mathématiquement linéaire selon un critère de profondeur visible des rides. RÉSULTATS: L'échelle à 12 points créée pour l'évaluation des pattes­d'oie sur les images ColorFace® s'est avérée réaliste, linéaire, robuste et précise pour les évaluations photographiques. La cohérence de l'échelle en termes de classement des images a été validée, ainsi que sa fiabilité et son acceptabilité en conditions réelles d'utilisation. En outre, les grades de rides obtenus par l'échelle SCAWA sont bien corrélés (R = 0,94) avec ceux obtenus pour les mêmes images par l'Atlas du vieillissement de la peau. La méthodologie de l'IA et le format numérique ont également apporté des résultats intéressants, tels qu'une meilleure harmonisation entre les experts et une représentativité plus élevée, c.­à­d. une diminution des images en dehors de la plage. CONCLUSION: L'échelle SCAWA tire le meilleur parti de l'apprentissage automatique pour fournir un outil numérique innovant visant à faciliter l'évaluation visuelle des rides sur les images, tout en optimisant les aspects de linéarité, d'homogénéité et de précision. Les avis enthousiastes des experts sur le format et la qualité de l'échelle sont prometteurs en ce qui concerne l'adaptation de la méthodologie à d'autres signes et une plus grande distribution de cet outil sur le marché de l'évaluation de l'efficacité des produits cosmétiques.

Comparative analysis of original face and skin-warped average face images for the scoring of skin attributes.

OBJECTIVES: Representative of a panel, an average face image could be used to analyse/display skin changes while alleviating image rights constraints. Therefore, we used landmark-based deformation (warping) of individual skin images onto their panel's average face, evaluating this approach's relevance and possible limits. METHODS: An average front face image was constructed from images of 71 Japanese women (50-60 years old). After warping individual skin images onto this average face, the resulting skin-warped average faces were presented to three experts who graded: forehead wrinkles, nasolabial fold, wrinkle of the corner of the lips, pore visibility and skin pigmentation homogeneity. Two experts estimated subjects' age. Results were compared to gradings performed on original images. RESULTS: Inter-expert grading shows excellent to good correlation whatever image type: from 0.918 (forehead wrinkles) to 0.693 (visibility of pores). Correlations between scoring of both image types are almost always higher than inter-expert correlations (maximum: 0.939 for forehead wrinkles-minimum: 0.677 for pore visibility). Frequencies of grades/ages are similar when scoring original and skin-warped average face images. Experts scores are similar in 90.6%-99.3% of the cases. Average deviations upon scoring both image types are smaller than average inter-expert deviations on original images. CONCLUSIONS: Scoring facial characteristics in original images and skin-warped average face images show an excellent agreement, even for perceived age, a complex feature. This opens the possibility of using this approach to grade facial skin features, monitor changes over time, and to valorise results on a face deprived of image rights.

Evaluation of a new connected portable camera for the analysis of skin microrelief and the assessment of the effect of skin moisturisers.

BACKGROUND: Silicone replicas and non-contact methods are effective methods to analyse the micrometric scale of the skin microrelief. Yet, they imply data capture in research facilities. The capabilities of a new connected portable camera were evaluated to analyse microrelief under nomadic conditions, also studying the effect of moisturisers. MATERIALS AND METHODS: 3D depth maps were constructed using shape-from-shading algorithms. Roughness heterogeneity (Spa) was computed, and skin profiles were extracted to calculate roughness amplitude (Ra, Rq), as well as furrows/plateaus characteristics. Validation of the connected camera was performed on tanned cowhide leather and on the inner forearm skin of a single subject. The forearms of 18 subjects (23-60 years old) were also evaluated. While living their regular life, they self-performed triplicate acquisitions at various times. The effects of a placebo and of cream containing moisturisers-saccharide isomerate, urea or xylitylglucoside-anhydroxylitol-xylitol-were investigated, using untreated control skin as a reference. RESULTS: Validation of the device on leather and forearm skin shows high repeatability. The 18 subjects show the known correlation between age and changes in microrelief. While testing formulas, 8 h after a single application, all decreased Spa (-1.6/-2.1 folds). Only saccharide isomerate and xylitylglucoside-anhydroxylitol-xylitol decreased Ra (-2.4/-2.8 folds). The sectional area of plateaus was reduced from -1.5 (urea) to -2.1 folds (xylitylglucoside-anhydroxylitol-xylitol). The height of plateaus is also decreased by all moisturisers, from -1.5 (urea) to -2.1 folds (xylitylglucoside-anhydroxylitol-xylitol). CONCLUSION: This novel camera device enables microrelief analysis under nomadic conditions, allowing monitoring its changes along the day and upon moisturisers' application.

Analysis of correlation and construction of a predictive model of skin transparency using parameters from digital images of the face.

BACKGROUND: Skin transparency is a cosmetic asset highly considered by Asian women. Resulting from complex light interactions within the skin, but still not fully understood, there is no simple method to measure it objectively. In this study, skin parameters from digital images were analysed to build a model predicting transparency. MATERIALS AND METHODS: Initially, 71 Japanese women (between ages 50 and 60 years) were recruited. This group was then extended to 262 women (between ages 21 and 60 years). Pictures of their faces were taken with the Colorface® under diffuse light and different polarisation angles. Experts graded their transparency using pictures. Pictures were also used to compute 958 skin colour and surface parameters from different regions of the face. RESULTS: In the initial group of 71 subjects, 109 parameters correlated with transparency. Half of them are from the cheek and relate to colour or colour homogeneity. If the cheek presented the largest proportion of correlated parameters, best correlations were usually found in other facial regions. Multiple regressions from some cheek parameters can predict up to 80% of transparency. Stepwise regression on parameters from 262 subjects led to a six-parameter model, which is highly correlated (R = 84.1%) with transparency. It combines skin texture, colour, colour homogeneity and gloss parameters. If half of them are from the cheek, the others are from the tear trough, the full face and the cheekbone. CONCLUSION: Using parameters from digital pictures exclusively, we propose a model that accurately reflects transparency. Including parameters previously shown to relate to transparency, this model should be useful for future dermatology and cosmetic research.

Differences between perceived age and chronological age in women: A multi-ethnic and multi-centre study.

OBJECTIVE: Accuracy in assessing age from facial cues is important in social perception given reports of strong negative correlations between perceived age and assessments of health and attractiveness. In a multi-ethnic and multi-centre study, we previously documented similar patterns of female facial age assessments across ethnicities, influenced by gender and ethnicity of assessors. METHODS: Here we extend these findings by examining differences between estimated age from digital portraits and chronological age (Δ age) for 180 women from three age groups (20-34, 35-49, 50-66 years) and five ethnicities (36 images of each ethnicity, assessed for age on a continuous scale by 120 female and male raters of each ethnicity). RESULTS: Across ethnicities, Δ age was smallest in French assessors and largest in South African assessors. Numerically, French women were judged oldest and Chinese women youngest relative to chronological age. In younger women, Δ age was larger than in middle-aged and older women. This effect was particularly evident when considering the interaction of women's age with assessor gender and ethnicity, independently and together, on Δ age. CONCLUSION: Collectively, our findings suggest that accuracy in assessments of female age from digital portraits depends on the chronological age and ethnicity of the photographed women and the ethnicity and gender of the assessor. We discuss the findings concerning ethnic variation in skin pigmentation and visible signs of ageing and comment on implications for cosmetic science.

OBJECTIF: La capacité à évaluer l'âge d'un visage avec exactitude en fonction de ses caractéristiques est important dans sa perception sociale. En effet, des corrélations négatives fortes ont été rapportées entre l'âge perçu d'un visage d'une part, et sa santé et attractivité d'autre part. Dans le cadre d'une étude multi-ethnique et multicentrique, nous avons déjà documenté, dans une démarche similaire, comment la perception de l'âge de visages féminins entre différentes populations, est influencée par le genre et l'origine des évaluateurs. METHODES: Ici nous approfondissons ces premiers résultats par l'étude des différences entre l'âge estimé sur portraits numériques de 180 femmes issues de 3 groupes d'âges (20-34, 35-49, 50-66 ans) et de 5 populations d'origine différente (36 images de chaque population) et leur âge réel (Δ âge), et ce par 120 évaluatrices et évaluateurs de chaque population évaluant l'âge des visages en utilisant une échelle continue. RESULTATS: Au sein des différentes populations d'évaluateurs, le Δ âge le plus faible a été trouvé chez les évaluateurs français et le plus élevé chez les évaluateurs sud-africains. Sur portraits numériques, les femmes françaises ont été perçues comme étant les plus âgées et les femmes chinoises les plus jeunes, par rapport à leur âge réel. Chez les femmes les plus jeunes, le Δ âge a été plus élevé que chez les femmes d'âge moyen et les plus âgées. Ceci a particulièrement été le cas lorsque l'on considère les interactions entre l'âge des femmes évaluées, et le genre et l'origine des évaluateurs, de façon indépendante ou liée, avec le Δ âge. CONCLUSION: Aux travers des différentes analyses, nos résultats suggèrent que l'exactitude avec laquelle l'âge des femmes est évalué sur images numériques de leur visage, dépend de l'âge réel et de l'origine de ces femmes photographiées, ainsi que de l'origine est du genre de l'évaluateur. Nous discutons ces résultats en regard des variations de pigmentation cutanée et de signes visibles de l'âge entre les différentes populations et commentons les implications possibles pour les sciences cosmétiques.

Cross-cultural perception of female facial appearance: A multi-ethnic and multi-centre study.

Humans extract and use information from the face in assessments of physical appearance. Previous research indicates high agreement about facial attractiveness within and between cultures. However, the use of a narrow age range for facial stimuli, limitations due to unidirectional cross-cultural comparisons, and technical challenges have prevented definitive conclusions about the universality of face perception. In the present study, we imaged the faces of women aged 20 to 69 years in five locations (China, France, India, Japan, and South Africa) and secured age, attractiveness, and health assessments on continuous scales (0-100) from female and male raters (20-66 years) within and across ethnicity. In total, 180 images (36 of each ethnicity) were assessed by 600 raters (120 of each ethnicity), recruited in study centres in the five locations. Linear mixed model analysis revealed main and interaction effects of assessor ethnicity, assessor gender, and photographed participant ("face") ethnicity on age, attractiveness, and health assessments. Thus, differences in judgments of female facial appearance depend on the ethnicity of the photographed person, the ethnicity of the assessor, and whether the assessor is female or male. Facial age assessments correlated negatively with attractiveness and health assessments. Collectively, these findings provide evidence of cross-cultural variation in assessments of age, and even more of attractiveness, and health, indicating plasticity in perception of female facial appearance across cultures, although the decline in attractiveness and health assessments with age is universally found.