3 ms·
Hi Rasmus, Cool work! I'm curious regarding the training dataset. What is the distribution of faces by race/age? Also regarding the raters, what is their distr
by fatjokes 11y ago
Hi Rasmus,
Cool work! I'm curious regarding the training dataset. What is the distribution of faces by race/age? Also regarding the raters, what is their distribution? (Race/age/cultural background)?
It's widely known that attractiveness is heavily dependent on cultural upbringing.
Signed,
Butthurt dev whose best pic only rated an "OK".
EDIT: You also rated Yoona (a Korean pop star) as just "Nice": http://imgur.com/uJVnQ9S http://imgur.com/uJVnQ9S. I guess that makes me feel better about my "OK". I'd stay out of Korea if I were you---I hear their fans aren't very forgiving.
- timofter 11y agoHi fatjokes, I am Radu, one of the authors. As mentioned in our article (http://arxiv.org/abs/1510.07867v1 http://arxiv.org/abs/1510.07867v1), the training dataset for attractiveness is from Blinq. The underage and above 37 years old face images were discarded. All people are heterosexual in our training dataset and mainly from Zurich area, Switzerland. Therefore, our model of attractiveness fits the cultural bias of Zurich, Switzerland. We consider faces.ethz.ch a little fun tool. I hope that the fans of Yoona will understand :) With more training data from abroad Switzerland our algorithms will fit their expectations. For age and gender we used much more and diverse data, therefore are more reliable for the majority of ethnic backgrounds.