Software start-ups getting big bucks to write code that can identify, find and link logos and brands in the billions of images posted daily. Larry Greenemeier reports
You post photos on social media sites for the enjoyment of your family and friends. But your snapshots are also a potential gold mine of information about what you spend money on for those sites and the companies that advertise on them.
Artificial intelligence software is on the horizon that can spot brands like Nike or Coke even in images without text or tags. Google, Facebook and other deep-pocketed investors are accelerating this software’s development. Even startup photo-sharing service Pinterest got in on the action earlier this year by buying the even more “start-uppy” VisualGraph, whose software could be used to find and link photos with similar content.
In general the software winning these investments uses machine vision, image recognition and/or visual search algorithms to identify objects and shapes as well as textures. One startup called Ditto Labs makes a search engine that specifically examines digital images for logos and brands.
So next time you’re chowing down in McDonald’s, smile for that selfie. You’re feeding that company’s bottom line and probably helping a tech startup earn its next million. On second thought, why are YOU smiling?
— Larry Greenemeier
[The above text is a transcript of this podcast.]