A S C R A P B O O K O F S O L U T I O N S F O R T H E P H O T O G R A P H E R
Enhancing the enjoyment of taking pictures with news that matters, features that entertain and images that delight. Published frequently.
2 May 2015
Visit How-Old.net to make the acquaintance of the HowOldRobot that may just be able to answer the eternal question of, well, how old you look.
The site uses Microsoft's Azure Face application programming interface to guess the age of any face it detects in a photo you upload.
Photos must be 3-MB or less and are not saved on the site.
HOW IT WORKS
In a recent Machine Learning blog post, Microsoft engineers Corom Thompson and Santosh Balasubramanian, told the story behind How-Old.net:
This may be hard to believe but it took a couple of developers just a day to put this whole solution together, starting with the pipeline from the Web page to the Machine Learning APIs to the real time streaming analytics and real time BI. This turned out to be a great example of the power of Azure services.
It's a three step job:
- Extracting the gender and age of the people in these pictures. "We found the ability of the face API to estimate age and gender to be particularly interesting and chose this aspect of it for our project," they point out.
- Obtaining real time insights on the data extracted above. "Once we have extracted the information we want from the uploaded pictures and web logs, we collect and analyze the data obtained from thousands of users uploading pictures to this site in real time."
- Creating real time dashboards to view the above results. "We use PowerBI to display the results in a real time dashboard," the wrote.
A TEST
We gave it a whirl -- and just loved it. It flattered us by suggesting we were 12 years younger than we look.
That was a full-on portrait (the one we use on this site) but we threw How-Old.net a curve and uploaded a group portrait from the Thanksgiving table with much smaller faces.
Going around the table the guesses were pretty inaccurate: -17 (our hostess), -15 (our better half), -50 (the matriarch), -2 (our 32 year-old neice), -8 (our host) and +15 (a 24-year old health nut).
If the artificial aging of the health nut isn't funny enough by itself, a little math reveals the parents of the health nut would have had him when his father was 14 and his mother 5 years old.
So Microsoft has a little work to do on facial recognition.
But accuracy isn't the only way it can win this game. Flattery will get you pretty far, too. Especially if you happened to be the matriarch at that table.
You can monitor the team's progress by following #HowOldRobot.