What Entrepreneurs Need to Know About Facial Recognition Technology
As the Fourth Industrial Revolution unfolds with billions of people sharing a wide and deep array of data — texts, tweets, GPS coordinates, all manner of photos, videos, environmental data, clickstreams, status updates, likes and reposts, pumping trillions of real-time signals into the digital universe — what does the future hold? This data is like food for the whale of artificial intelligence.
In terms of a resource, this AI-food-rich data ocean makes the California gold rush, or the Texas oil boom, seem like tiny puddles. Vast amounts of data are flooding the digital space on a global level. AI-based algorithms will be propelling innovation in every sales arena from products to services and the more data you have, the more accurate the algorithm. Collecting and processing “big data” has become a focus for companies large and small.
And how does the AI whale digest this data? Through interconnected devices with embedded “eyes.” Termed “deep learning,” these artificial neural networks use layered machine learning algorithms that mimic the structure of animal brains. Utilizing gigantic data pools, deep learning can identify and interpret complex patterns much in the same way as the human brain. Some artificial neural networks are now extremely adept at employing these patterns to mimic the way humans recognize faces.
So, which company is in front of the deep learning fleet? Facebook, of course. Facebook holds the single largest collection of facial data, and in 2015 it introduced a greatly enhanced version of its “tag photos” feature, DeepFace, which employs a nine-layer neural network that matches features in separate photographs with 97.25 percent accuracy. DeepFace not only connects your face with your name, but it can literally pick your face out of a crowd, and a human brain is only .28 percent better at this than the program. Facebook has invested big-time in DeepFace, spending billions of dollars devouring the competition (including Face.com, Masquerade and Faciometrics).
Recently Facebook was granted a new patent, “Techniques for emotion detection and content delivery,” which captures users’ facial expressions via the camera in real time as they scroll through their feed, tracking their emotions when exposed to various content. This emotional data could not only personalize your Facebook feed at a whole new level, but could also link to live in-store cameras, matching and identifying shoppers, calling up information gleaned from Facebook and identifying the shopper’s current moods.
