Google Launches MediaPipe for Raspberry Pi, Offering a Python SDK for Simplified On-Device ML
Examples include on-device audio, text, and image classification, object detection, gesture recognition, and facial landmarking.
Google has announced the launch of MediaPipe for Raspberry Pi, offering a Python-based software development kit (SDK) for machine learning (ML) tasks complete with examples for audio classification, text classification, gesture recognition, and more.
"Back in May we released MediaPipe Solutions, a set of tools for no-code and low-code solutions to common on-device machine learning tasks, for Android, web, and Python," Google's Paul Ruiz, developer relations engineer, writes of the release. "Today we’re happy to announce that the initial version of the iOS SDK, plus an update for the Python SDK to support the Raspberry Pi, are available."
MediaPipe Solutions, first unveiled in preview form in December last year before being launched properly at Google I/O in May, is designed to offer developers a leg-up in on-device machine learning work. In its most recent update, the Python side of the platform has gained official support for the Raspberry Pi range of single-board computers — though performance will vary from device to device, with peak performance achievable only on the Raspberry Pi 4 and Raspberry Pi 400 models.
"Aside from setting up your Raspberry Pi hardware with a camera, you can start by installing the MediaPipe dependency, along with OpenCV and NumPy if you don’t have them already," Ruiz offers as part of a quick-start guide to using MediaPipe on a Raspberry Pi. "From there you can create a new Python file and add your imports to the top. You will also want to make sure you have an [ML] model stored locally on your Raspberry Pi."
At the time of writing, Google had published Raspberry Pi-compatible MediaPipe Python examples for audio classification, facial landmarking, gesture recognition, image classification, object detection, and text classification; other examples written for the earlier generic Python release have not yet been marked as compatible.
More information on MediaPipe is available on the project webpage, with the examples published to GitHub under the permissive Apache 2.0 license.