Amazon Starts Shipping Its DeepLens Camera For Developers


In its re:Invent keynote event last year, Amazon talked about launching its AWS DeepLens camera which is similar to Google’s Clips camera and uses Artificial Intelligence to take better shots. The only difference between the two, however, is that DeepLens is specifically targeted towards the developers.

A few months ago, Amazon has even taken pre-orders for the product and now it has started shipping the camera to the developers. According to the company DeepLens is designed to allow developers up to speed with Amazon’s various forays into AI, Internet of Things (IoT) and will also further server less computing. Amazon further hopes that DeepLens will convince developers to create apps using its services like the AWS Greengrass and also AWS Lambda.

The AWS DeepLens is basically a HD video camera with on-board computing and is optimized for deep learning. It also integrates well with Amazon SageMaker and AWS Lambda. Amazon claims that the device will allow developers with their first inference in less than ten minutes. It also comes with pre-built models, tutorials, demos and even some examples.

The device supports all sorts of connectivity on-board including two USB ports, a micro HDMI together with dual-band Wi-Fi. It has an Intel Atom 8 processor at its heart and 8GB RAM and runs on Ubuntu. The camera is 4GB and supports shooting in 1080p (Not so great but perfectly adequate for most uses).

The best part is that it is loaded with pre-trained models that support image recognition of different models including cats and dogs and even inanimate objects and then there is also an included model that will be able to identify about 30 different actions like playing a guitar, for example. It also offers an online template that will help in leveraging these technologies. Further, the team working on DeepLens is planning to add a model for tracking head poses

Practically, developers will not be able to do a lot of the development on DeepLens hardware itself as it is essentially a small computer. What could probably be better for them is to use a more powerful computer and then deploy DeepLens with the help of AWS Console.

Amazon’s VP for AI Swami Sivasubramanian said, “The whole rationale behind DeepLens came from a simple question that we asked ourselves: How do we put machine learning in the hands of every developer. To that end, we brainstormed a number of ideas and the most promising idea was actually that developers love to build solutions as hands-on fashion on devices.”