In this interview I spoke with Richard Carriere GM of Cyberlink America on using Face Recognition in Retail.
Founded in 1996, CyberLink Corp. (5203.TW) is a world leader in multimedia software and AI facial recognition technology. The company owns over 200 patented technologies by which it pushes the evnelope for the multimedia experience and other technologies including Facial Recognition with the newly launched FaceMe product.
In this interview Darius covers how CyberLink came up with the idea for the products and how it plans to apply and implement it in retail applications.
DV. What is CyberLink and your role in driving and running the company?
RC. I am responsible for US business and Global Marketing for the company. CyberLink is in the market for around the last 25 years focusing on digital media software and apps, basically anything related to viewing, editing, recordings, streaming your content, etc. If you have used a PC/Laptop in life you have most likely used one of the apps built by us.
We developed the face recognition technology about 4 years ago when we were transferring one of our apps into a smartphone for a product that does virtual makeup. The app YouCam Makeup has been downloaded nearly 1 Billion times.
The app needed to have AI which recognizes faces by not just using data from the cloud, like most other tools in the market, but should be quick enough to detect the face even without using the internet. This technology has been tested with multiple organizations including POC's in retail.
DV. Basically you productized the technology into an SDK and it can be applied to multiple verticals?
RC. Yes. FaceMe technology can be applied to multiple verticals because we developed it on smartphones already for Android and iOS. We also developed it to run on Linux and Windows. So basically we run across any platform. We also have different versions of our algorithms and some of them can be very precise like 99.8% precision.
DV. What problems specifically it is solving for retailers?
RC. Actually there are 3 or 4 layers of problems or opportunities that can be addressed and they revolved around digital signage, around customer identification loyalty, around security.
If we talk about just the first level i.e. digital signage or statistic collection so you can use the level just of identify the statics like who is walking out of the store or walking at a sign or display and by reading vectors on the face. It's pretty good to identify if this person is a man or a woman, how old are they and their facial expressions. With that, we can collect very wealthy information that could be used to track an enhanced experience in a store.
Q. Are people okay to be recorded or viewed on camera in stores? Have you done a study about that?
Ans. Well, there are different things. Everybody walks into different stores so they are under cameras for security purposes. Probably more times then they would realize.
So in that sense, that level of usage of our technology doesn't even need to store or identify uniquely every visitor. There is a possibility to use it simply compile statistics.
No data is kept, so legally there is no problem doing that. Especially stores that are private entities.
So that’s pretty much anywhere in the world.
There are some cases where people would need to see a sign, even for security nowadays.
That says there are close capture cameras in operation but its no different that than.
That’s what they can do.
Of course, we can do more than that.
Another use case that could and doesn’t need to record details or keep a record of who is viewing at the camera is digital signs.
So based on who is looking at it.
Let’s say you are in a self-serving kiosk at a fast-food restaurant.
You are a teenager at 4 o’clock in the afternoon, it might be a content management system that will push to you menu options that are a little bit different than if you are a nice grandmother coming at 10 Am to drink a cup of tea.
So that’s something that could happen and by the way even if we work to collect information the technology identify vector points on the face it doesn’t even have to take a picture of the people and this is highly encrypted so there is nothing that can be done with the data unless its part of that system, so far there is nothing too scary about it and if anything enhances the customer experience and value to the retailer.
So that’s what I can say as the first level.
Now there is a second level, the moment we have exclusive content being developed for some luxurious goods retailer.
It's being able to identify self optident VIP customers.
So imagine you a VIP customer at a luxury brand store and you go to the store on Madison Avenue in New York. Using FaceMe technology the store can offer highly personalized recommendations to the customer as they enter and browse the store.
This is an overview of a small part of the interview, make sure to listed to the complete Interview.
Check out CyberLink's many products at CyberLink.com and you can connect with Richard on Linkedin.