Microsoft’s Project Oxford helps MyMoustache.net gather Movember data from 145 countries

Microsoft’s Project Oxford has proven an impressive machine learning tool with various applications. And it has powerful potential as Microsoft has introduced machine learning APIs so developers can integrate Project Oxford’s capabilities into their platforms. These changes have included the addition of “emotion, video, spell check, speaker recognition, the Custom Recognition Intelligent Service and improvements to our existing Face APIs.”
And of course, Microsoft’s machine learning team has taken multiple opportunities to create fun showcase sites using Project Oxford such as How-old.net and TwinsOrNot.net. Most recently, the team created MyMoustache.net, which uses Project Oxford’s Face API to detect and quantify facial hair to again showcase the technology, but also to help raise awareness for Movember and the Movember Foundation. The website works by allowing the user to upload a photo of their facial hair, and having machine learning algorithms rate your stache and compare it to the rest of the world.
Now that Movember is officially over, the machine learning team has posted the results of all the submissions, and there was quite a global response to MyMoustache.net. In total, 145 countries participated. The continent with the highest average mustache length was North America, but when broken down country by country, the island nation of Mauritius topped the charts with the greatest average moustache length. Of all the participants, 27% received the “ultimate mo” status for achieving a perfect score for moustache length. To see more of the results, you can read aboutMyMoustache.net and Movember here.
Stache-4.JPG-550x0 Microsoft's Project Oxford helps MyMoustache.net gather Movember data from 145 countries
Via TechNet.com
Stache-4.JPG-550x0 Microsoft's Project Oxford helps MyMoustache.net gather Movember data from 145 countries
Via TechNet.com
For all the fun with showcase sites like MyMoustache.net, there was some serious computer science behind it. On technet.com, the Machine Learning team writes:
Jokes aside, one might ask what it takes to train a new ML algorithm, specifically one that measures facial hair? We took it as a classification problem. Face hair normally grows on three face areas: moustache above lips, beard under lips, and sideburns in front of ears. “Density” is introduced to describe how bushy that facial hair is in those face areas – level ranges include clean-shaven (no facial hair) stubble, medium and bushy.
If you missed your chance to participate in Movember, fret not. MyMoustache.net is still up and raising money for men’s health through collaboration with the Movember Foundation.

0 comments