Faces have a lot to tell about a person’s health, but it comes as a bit of a surprise that there is a way to estimate a person’s body mass index (BMI) by simply analyzing various characteristics in a mug shot.
Researchers at West Virginia University, Morgantown published a study in Image and Vision Computing describing the development of a computer vision algorithm that does just that. The software compares various distances against each other, such as the length of the face vs. cheek width, and spits out the BMI. They tested the algorithm on over 14,000 faces with known BMIs and showed that the approach is fairly effective and can be improved to produce more effective results. Besides being used for medical applications, the researchers envision the same technology being applied to help you parse through potential mates on online dating sites.
Study in Image and Vision Computing: A computational approach to body mass index prediction from face images…
(hat tip: NewScientist)