Open AccessMeasuring asymmetry from high-density 3D surface scans: An application to human faces
Author(s) -
Omid Ekrami,
Peter Claes,
Julie D. White,
Arslan Zaidi,
Mark D. Shriver,
Stefan Van Dongen
Publication year - 2018
Publication title -
plos one
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.99
H-Index - 332
ISSN - 1932-6203
DOI - 10.1371/journal.pone.0207895
Subject(s) - asymmetry , fluctuating asymmetry , landmark , computer science , iterative closest point , artificial intelligence , facial symmetry , computer vision , pattern recognition (psychology) , physics , biology , orthodontics , medicine , quantum mechanics , evolutionary biology , point cloud
Perfect bilateral symmetry is the optimal outcome of the development of bilateral traits in the absence of developmental perturbations. Any random perturbation in this perfect symmetrical state is called Fluctuating Asymmetry (FA). Many studies have been conducted on FA as an indicator of Developmental Instability (DI) and its possible link with stress and individual quality in general and with attractiveness, health and level of masculinity or femininity in humans. Most human studies of facial asymmetry use 2D pictures and a limited number of landmarks. We developed a protocol to utilize high-density 3D scans of human faces to measure the level of asymmetry. A completely symmetric spatially dense anthropometric mask with paired vertices is non-rigidly mapped on target faces using an Iterative Closest Point (ICP) registration algorithm. A set of 19 manually indicated landmarks were used to validate the mapping precision. The protocol’s accuracy in FA calculation is assessed, and results show that a spatially dense approach is more accurate. In addition, it generates an integrated asymmetry estimate across the entire face. Finally, the automatic nature of the protocol provides a great advantage by omitting the tedious step of manual landmark indication on the biological structure of interest.