Deep neural networks were used to extract features. have more feminine facial features, while lesbian facial features are more masculine.
At the univariate level, lesbian and heterosexual women differed in 17 facial features (out of 63) and four were unique multivariate predictors in logistic.
The algorithm took into account both 'fixed' and 'transient' facial features ('Deep neural networks are more accurate than humans at .. in American culture, heterosexual men and lesbians tended to wear baseball caps.".