We’ve all heard the timeless saying: Age is but a number. We’ve also all experienced the surprisingly large amount of lying that’s associated with the number. What may be inconsequential to a toddler at their birthday party is a closely guarded state secret to a veritable pensioner.
As a result, with the passing of many seasons, it becomes exponentially harder to ask a person for their age, leading you to guess the age you want the baker to write on the chocolate cake with vanilla icing. Unfortunately, it may result in copious amounts of cake being splattered all over your person by the furious birthday celebrator due to the fact that number now plastered on your face is few years to old.
In these difficult times that are rampant with flying cakes, Microsoft (Yes, the same company trying to desperately move people away from Windows XP) have recently released a tool for facial recognition which an internal team decided to take a further step with and build a website. Appropriately named how-old.net the website uses the same facial recognition technology to predict the age of someone using a picture of themselves (which will invariably be a selfie.)
I immediately jumped onto said website and started uploading pictures of myself and various celebrities. Sadly however the age guessing website only managed to guess the correct age for Jared Leto; Robert Downey Jr. and Angelina Jolie were both estimated to be in their 70s and for some odd reason the website’s facial recognition tool couldn’t seem to recognise my face. Whatever. (Funnily enough a picture of my cat churned out the age of 36, must have been all the facial hair?)
Jokes and fun aside, the surprisingly fun website is testament to how powerful machine learning technology (the precursor step to A.I) can be. Technically speaking the tool is based on various algorithms and is designed to learn from its mistakes and from inputs of its users so that it becomes more accurate.
Of course this isn’t the first time facial recognition software has been used, Facebook uses a similar tool to help you tag friends in case you forget the name of that person you took a selfie with. The limitations however do not end there: experts are already trying to implement smart facial recognition tools into daily applications such as criminal investigations and the like.
You can test out the tool at how-old.net
