Blood relationship revealed by smile

In 2012, some four hundred people contributed to the most comprehensive smile database ever. UvA researchers recorded the laughter of hundreds of NEMO visitors. They were able to do this thanks to the Science Live program in which young and old can participate in real research. By means of the UVA research we already knew to what extent a smile is genuine or not, but now it also appears that it is possible to reveal blood relationship by analyzing your smile. The Science Live program takes place in NEMO and is made possible in part thanks to the support of KNAW and NWO.

Theo Gevers, Hamdi Dibeklioglu and Albert Ali Salah, researchers at the University of Amsterdam, developed software for automatic determination of the blood relationship between people based on their smiles.

During the research in NEMO, videos were made of hundreds of visitors resulting in the largest smile dataset in the world: UvA-NEMO Smile Database. The camera also recorded other expressions such as anger, surprise and fear. Dibeklioglu and Gevers show that the dynamics of a smile (the way in which facial muscles move, such as intensity and acceleration) yield a wealth of information with which it can be accurately determined whether someone really laughs, and whether there is blood relationship between people. Even if people look similar to each other, the way they laugh will reveal the kinship. It is the first time that computer software can accurately recognize the relationship between people, such as mother-son relationship, father-daughter, brother-brother, sister-sister, based on facial expressions. Further research is being conducted into other expressions such as anger, surprise and fear of detecting blood relationship.
Earlier, the research by Gevers, Dibeklioglu and Marcel Lucassen showed that as a woman you look younger when you laugh, but only when you are over forty. Women under the age of 40 can look better neutral if they want to appear younger. Furthermore, the color of the light source has an influence on age estimation. Through this research, there is now more insight into what exactly happens to your face when you are genuinely smiling, and how that affects your external (apparent) age. With this software is developed for automatically calculating the age of people with an average error of 4.5 years. If people estimate the age of other people, accuracy is generally lower (average error around 7 years). Furthermore, software has been developed for the automatic determination of viewing direction and head movements. The database is publicly available. A preview can be seen at www.uva-nemo.org.
Theo Gevers has been researching digital facial recognition for several years now. Gevers received a Vici grant from NWO's Innovational Research Incentives Scheme for his pioneering work in 2007.

About Science Live

With the Science Live research program, NEMO brings science and the public together. This not only provides valuable data, but also gives the public an interesting insight into the scientific kitchen. First of all, the visitor meets a scientist 'live' and experiences what it is to take part in an investigation. Science Live is loved by visitors to NEMO and by scientists. NEMO would like to get in contact with even more scientists who want to conduct their research via Science Live. More information about Science Live at: www.sciencelive.nl. Science Live comes about through collaboration with NWO and the KNAW.

Video: 10 Facts Proving Even the Queen Has Strange Habits

Leave Your Comment