​Face Recognition: Marketing Made Easy

I was watching the movie “Guardians of the Galaxy”, and the scene with the little raccoon, named Rocket, was using his little tablet to look for people with bounty on their head.

It got me thinking, “I wonder if an app like this exists?”

A quick search revealed an underwhelming short list mostly for locking/unlocking the device.

I found an Android app (ncit.face.detection.recognition), sadly it didn’t work that well (at least on my Samsung Note 4 and Note 8). This was a bit disappointing, so I did a few Google searches and found a sweet little gem called Opencv.

Get notified.

Join the thousands of marketers, industry professionals and intrigued readersand get notified when a new post goes live!

Right off the bat, the Opencv framework included a sample code for face recognition.

"Well,” I thought. “I’m on the right track.”

With bit of extra coding and a bit of user interface work, I had a running version for real-time recognition and training for detected faces. This was awesome, and it’s free.

Unfortunately, training the application means I have to write all the information myself, which is not exactly ideal. So, why not use the images already provided by my friends on social media? Yes, I know, a little creepy, but this is science we are talking about people. Science!

The good (or bad, depends on who you are) about social outlets is relevant information is right there, upfront and without too much digging. My early prototype was encouraging, since it was able to identify most of my friends and show their name, and latest status updates. For someone who easily forgets faces, this is freakishly cool app.

But what about for marketers? We already have automation systems connecting users actions to user profiles in order to measure ROI on an entire campaign, ad or piece of content.

Some systems even work to consolidate all this data into a single user profile, letting us know everything about the user and their habits. But, this takes time, and a lot of collateral upfront (emails, forms, database configurations, etc.) to ensure success.

If a system could recognize a person immediately by the picture on a social profile, and then associate all known data to that person, the idea of marketing automation completely changes. Facebook is one of the leaders in this technology, as you may know from when you tag friends, relatives or others in images. But to have your own system digitally implemented to recognize current, prospective or interested customers, albeit creepy, very powerful in the world of marketing.

Even though the world of face recognition software is quite a bit away from consumer/business use, my experiment and learning lesson yielded some positive results. We will keep experimenting and testing to see how exactly a system like this can be used for marketing purposes in the future.