Amazon DeepLens, a machine learning-enabled video camera, is now available for all developers at US$249. Amazon claimed that the DeepLens is able to run machine learning models via TensorFlow and Caffe in less than 10 minutes for developers.
First announced at Amazon’s re:Invent conference last November, the DeepLens is a camera tailored for developers who want to build vision-centric machine learning prototypes.
The device supports computer vision models using Gluon, SageMaker imports and tools to optimize models, as well as Amazon’s Kenesis video streams. It’s also deeply integrated with Amazon’s AWS IoT service, Greengrass and other AWS tools.
“DeepLens allow developers to hone everyday skills and apply beyond in their organizations. With machine learning you need a lot of data. A video camera can capture things inside house, office and categorize them,” said Dr. Matt Wood, general manager for machine learning services at Amazon, in an interview with ZDNet.
Project templates provided by Amazon include an object detection model that can distinguish up to 20 objects, a face detection model, a motion differentiation model that can tell 30 different actions like playing guitar and eating, and a style transfer model to render the camera image into the style of modern painting.
DeepLens aims to let developers, who either have knowledge of machine learning or not, explore the possibilities with intelligent visual devices.
Since most developers, or all developers, like to make something with devices, Amazon thought it could just create a machine learning-powered device shipping to mass developers.
DeepLens features a 4 megapixel camera with 1080p video, 2D microphone array, Intel Atom processor and 8GB of memory. It runs on Ubuntu 16.04 system. It’s available on Amazon at US$249.