Google Brings AI, Voice and Imaging to Raspberry Pi

The introduction of Google's do-it-yourself kit has a gentle introduction to learning about artificial intelligence and Google has already come out with two separate projects to support this idea. The first one, known as the AIY Voice Kit is a voice recognition kit for Raspberry Pi. It includes the following: same VoiceHAT (Hardware Accessory on Top), mic board, speaker, components, connectors and cardboard form. To add depth to the technology, the Google Assistant SDK is configured to bring hot word detection, voice control, natural language learning, Google smarts and many more to the kit. It's not restricted to just that, however as the user is able to extend the voice kit by implementing additional vocabularies using TensorFlow.

The second project, AIY Vision Kit is the newest and latest addition to Raspberry Pi. The newest kit features on device neural network acceleration that allows powerful computer vision without the use or storage of cloud connections.  In order to use it, the user must add a Raspberry Pi, Raspberry Pi Camera, an SD Card and a power supply which must be purchased on their own. The kit features a cardboard outer shell, VisionBonnet circuit board, RGB arcade-style button, piezo speaker, macro/wide lens kit, tripod mounting nut and other connections for the kit.


AIY Vision Kit parts (Photo Credit: Google)

The main part of the kit is the VisionBonnet board for Raspberry Pi which has the Intel Movidius MA2450, a low-power vision processing unit designed to run neural network modes on the device.

The software comes with three TensorFlow neural network models for different uses and implementations. One of the models has the ability to recognize a thousand different objects, the second model has the ability to recognize faces and the expressions they may hold while the third model can spot differences between a person, cat or dog.


VisionBonnet for Raspberry Pi (Photo Credit: Google)

The kit also features a Python API that allows users to change RGB colors, adjust piezo element sounds and access four GPIO pins.

Google has also offered more articulate guidelines while using the kit and maintaining much easier access for the one using it. The implementation and idea here are to give it a wide range of ideas for the younger audience, as well. The Android App also provides some strong reference points to help get started and progress the kit for those wishing to get started on it.

Both products will be aiming to land in retail stores sometime in April and other stores around the world. The prices for both products isn't cheap, however, as the Vision pack comes at $90 and the voice pack at around $50 which is a fair price considering what you'll be getting in the kits.

Have a story tip? Message me at: cabe(at)element14(dot)com

http://twitter.com/Cabe_Atwell