The relevance and accuracy of an AI algorithm-based descriptor on 23 facial attributes in a diverse female US population

BACKGROUND

The response of AI in situations that mimic real life scenarios is poorly explored in populations of high diversity.

OBJECTIVE

To assess the accuracy and validate the relevance of an automated, algorithm-based analysis geared toward facial attributes devoted to the adornment routines of women.

METHODS

In a cross-sectional study, two diversified groups presenting similar distributions such as age, ancestry, skin phototype, and geographical location was created from the selfie images of 1041 female in a US population. 521 images were analyzed as part of a new training dataset aimed to improve the original algorithm and 520 were aimed to validate the performance of the AI. From a total 23 facial attributes (16 continuous and 7 categorical), all images were analyzed by 24 make-up experts and by the automated descriptor tool.

RESULTS

For all facial attributes, the new and the original automated tool both surpassed the grading of the experts on a diverse population of women. For the 16 continuous attributes, the gradings obtained by the new system strongly correlated with the assessment made by make-up experts (r ≥ 0.80; p < 0.0001) and supported by a low error rate. For the seven categorical attributes, the overall accuracy of the AI-facial descriptor was improved via enrichment of the training dataset. However, some weaker performance in spotting specific facial attributes were noted.

CONCLUSION

In conclusion, the AI-automatic facial descriptor tool was deemed accurate for analysis of facial attributes for diverse women although some skin complexion, eye color, and hair features required some further finetuning.

Developing an Artificial Intelligence (A.I)-based descriptor of facial appearance that fits with the assessments of makeup experts

OBJECTIVE

To develop an A.I-based automatic descriptor that detects and grades, from selfie pictures, 23 facial signs, hairs included, as a help to making-up procedures.

MATERIAL AND METHODS

The selfie images taken in very different conditions by 3326 women and men were used to create (90% of dataset) and validate (10% of dataset) a new algorithm architecture to appraise and grade 23 different facial signs such as lips, nose, eye color, eyebrows, eyelashes, and hair color as defined by makeup artists. Each selfie image was annotated by 12 experts and defined references to train Artificial Intelligence (A.I)-based algorithm.

RESULTS

As some the 23 signs present a continuous or discontinuous feature, these were analyzed by two different statistical approaches. The results provided by the automatic descriptor system were not only in good agreement with the expert's assessments but were even found of a better precision and reproducibility. This automatic descriptor system has proven a good and robust accuracy despite the very variable conditions in the acquisition of selfie pictures.

CONCLUSION

Such automatic descriptor system seems providing a valuable help in making-up procedures and may extend to other activities such as Skincare or Haircare. As such it should allow large investigations to better evaluate the consumers' needs of esthetical improvements.

Changes in the eye contour signs due to age among Mexican women: Comparison with women of other ethnic origins

OBJECTIVE

To evaluate the changes in certain ocular signs because of age, among Mexican women and to compare these with those previously obtained on women of other ethnicities.

MATERIAL AND METHODS

Photographs were taken of the faces of 203 Mexican women of different ages, under standardized conditions. These photographs allowed us to focus and define nine ocular signs, which were then graded by 15 experts and dermatologists, using standardized scales provided by a reference Skin Aging Atlas. Hence, the same protocol, previously used on 3240 women of four other ethnic origins (Caucasian: 600; Chinese: 990; Japanese: 1010; Indian: 300; African: 340), allowed us to compare the results obtained on Mexican women.

RESULTS

The severity of crow's feet wrinkles for Mexican women shows a more intense progression with age than those of women from other ancestries. The lower eyelid wrinkles of Mexican women, that show weak or erratic changes with time, were however found to have higher and significant severity at a young age. Pigmentation disorders of their ocular area were found less pronounced as compared to those of Asian or Indian women. Eye bags were observed with similar slight changes in all five ancestries.

CONCLUSION

The present work confirms that the clinical signs of age for the female eye area develop differently depending on the ethnic origins of the volunteer.