IEEE Access (Jan 2022)

A Comparative Study on the Perceived Attractiveness of Oriental and South Asian Faces

  • Muhammad F. Mughal,
  • Xiaoxuan Liu,
  • Ming Ronnier Luo,
  • Michael Pointer

DOI
https://doi.org/10.1109/ACCESS.2022.3205311
Journal volume & issue
Vol. 10
pp. 106415 – 106421

Abstract

Read online

The computation of perceived attractiveness from facial images has long been a research topic. Many models have been developed to predict the attractiveness of the face from the individual models that have been used to describe the geometry of the face (symmetry, golden ratio and neoclassical canons, according to artists from the Middle Ages, and a combination of the three). An experiment was conducted based on Oriental and South Asian ethnic groups, represented by Chinese and Pakistani facial images. Visual assessments of perceived attractiveness were carried out using a 6-point categorical scale and the results were used to derive a new set of facial feature ratios that maximized the perceived attractiveness of the two ethnic groups. The results were also used to develop a new polynomial model of attractiveness, and to test four existing models. The new model performed the best for Oriental faces. The new model was also best for South Asian faces together with the combined model. Ethnic group differences did not have a significant impact on the perceived attractiveness of the two groups. A set of new facial ratios for the two ethnic groups was determined to maximise attractiveness.

Keywords