AI on the Edge
summary
Demonstrates building an offline image classifier exported from Custom Vision and Lobe and deployed as a Docker module on Azure IoT Edge that runs on a Raspberry Pi using an OpenCV camera, sends local predictions up to IoT Hub and speaks labels with Bing Speech. Shows transfer learning from public datasets (Fruit365 ~50,000 images), exports a compact TensorFlow Lite model (~2–5 MB), and mentions ONNX/Core ML for cross‑platform use; demos mostly run live with a few hiccups but the Pi classifier performs in real time.
AI-written summary · about
chapters (6)
- 0:00Start
- 5:45Common Deep Neural Networks
- 6:23Computer Vision Tasks
- 20:42Azure lot Hub in the Cloud and on the Edge
- 25:57Azure loT Edge Design Principles
- 31:33lot Edge Runtime Concepts
related talks
45mFrom Data to AI on The Edge
43mIntersection of Awesome: Building AI enabled IoT solutions powered by the cloud
49mEnable IoT with Edge Computing and Machine Learning
41mAI + Machine Learning for .NET Developers
01h 00mLearn Live: Get started with AI on Azure
59mML and the IoT: Living on the Edge
36mAI-Powered Real-Time Quality Control
01h 01mML and the IoT: Living on the Edge