Amazon’s Cloud-Hosted AI System Can Recognize People and Things

Amazon has announced a new AI system based on deep learning that can not only recognize faces, but the emotions they express, as well as objects, animals, and all kinds of other things. It’s called Amazon Rekognition.

The system is available through Amazon Web Services, having launched in Northern Virginia, Oregon, and Ireland. Developers can access it for free right now, using its API functions to set up their own facial and object recognition systems, and will only have to pay for use on a tiered basis once their platforms need to analyze more than 5,000 images per month or store over a thousand recognized vectors (the minutiae upon which the system’s recognition is based) per year.

In a blog post announcing the system, AWS’s Jeff Barr offers a few examples of its applications: It could be used to “compare a face on a webcam to a badge photo” for access control, to scan photos “for objects or people of interest or concern,” or even to “build ‘smart’ marketing billboards that collect demographic data about viewers.”

There are other technologies that already do this kind of thing, but they tend to be specialized solutions such as Safran’s surveillance feed scanning system, or NEC’s biometric retail offering. Amazon seems to be attempting to disrupt those models with a cloud-hosted system that can be adapted to a wide range of applications, and one that could prove to be a valuable tool for a number of organizations.