Amazon just launched a camera that thinks.
The AWS DeepLens AI Deep Learning Video Camera, also called the DeepLens, was introduced in November during the Amazon Web Services (AWS) re:Invent conference. Now it’s finally available for for $249 after much delay (it was supposed to begin shipping by “early 2018”).
DeepLens is similar to the AI-powered Google Clips camera, but while Clips is targeted at consumers, DeepLens is a new toy for developers.
According to Amazon’s website, it’s the first video camera designed to teach deep learning basics and optimized to run machine learning models on the camera. This kind of machine learning is usually done by gathering information on one device and computing in the cloud — as opposed to doing it all on one gadget.
It ultimately helps people create their own deep learning tools and has six sample projects built into it: object detection, hot dog not hot dog (hey Silicon Valley, your sitcom has fans at Amazon), cat and dog, artistic style transfer, activity detection, and face detection.
The camera is currently optimized to learn from datasets available in Apache MXNet, but those in TensorFlow and Caffe will soon be compatible with it as well.
The camera runs on the free, open-source operating system Ubuntu and an Intel Atom X5 processor with four cores and four threads. It does not currently work with Alexa but is equipped with 8GB of RAM, 16GB of storage, microphones, a micro HDMI port, two USB ports, a speaker, headphone jack, a 4-megapixel (1080p video) camera and an Intel ninth-generation graphics engine.
DeepLens has a camera comparable to that of a webcam, but its operating capabilities are basically as powerful as that of a computer. DeepLens currently only ships within the United States.