Amazon’s newest hardware product is meant to help developers run deep learning and image recognition models directly from a video camera.
At its annual re:Invent conference in Las Vegas, Amazon Web Services today unveiled DeepLens, a $249 wireless programmable HD video camera that can capture 1080p video and audio and run deep learning models for image detection and recognition — all on the device itself.
Matt Wood, general manager of artificial intelligence at AWS, showed off the new device on stage and noted that it was “specifically built to help developers hone their skills with machine learning.” DeepLens, powered by an Intel Atom X5 processor, can accurately detect and recognize objects, like classifying food or animals, for example. It also enables face recognition and can identify activity like brushing teeth or applying lipstick.
Wood did a quick demo that showed how DeepLens works. He held a physical album up to the camera and smiled — it immediately recognized that it was Pink Floyd’s The Dark Side of the Moon, in addition to the fact that Wood’s face had a “positive expression.”
Developers can train models in Sagemaker, a new machine learning service from AWS, and deliver them to the DeepLens device from the AWS Management Console, Wood said.
AWS also announced a new machine-learning training system called SageMaker, designed to make it easier for customers interested in using machine learning within their applications to get sophisticated models up and running. Customers can train their own models based on various criteria, or they can pick from models prebuilt by AWS.
“You’re really experimenting with machine learning, and that’s the best way to refine your skills,” said Matt Wood, general manager of deep learning and AI for AWS, during the keynote.
Developers can pre-order DeepLens here; Amazon will ship by “early 2018.”
“All of this hardware, software, and data come together to make the AWS DeepLens a prime example of an edge device,” AWS noted in a blog post. “With eyes, ears, and a fairly powerful brain that are all located out in the field and close to the action, it can run incoming video and audio through on-board deep learning models quickly and with low latency, making use of the cloud for more compute-intensive higher-level processing.”
The product is similar to something Google unveiled last month with “Google Clips,” an AI-powered hands-free camera, though DeepLens seems to be more targeted at developers.
This isn’t the first time Amazon’s cloud computing business has unveiled hardware at its annual conference. Last year, it debuted AWS Snowmobile, a semi-trailer that will come to your facility and literally pick up an exabyte of data.